27 RMM Winter Edition 2022 Solutions

Number 27 WINTER 2022
R M M
ROM A N IA N MAT HEMAT IC AL MAG AZINE
SOLUTIONS
Founding Editor
DANIEL SITARU
Available online ISSN-L 2501-0099
www.ssmrmh.ro
www.ssmrmh.ro
Proposed by
Daniel Sitaru – Romania
Mehmet Şahin-Ankara-Turkyie
Marian Ursărescu-Romania
George Apostolopoulos-Messolonghi-Greece
Ionuț Florin Voinea-Romania, D.M. Bătinețu-Giurgiu-Romania
Cristian Miu-Romania, Neculai Stanciu-Romania
Florică Anastase-Romania
Vasile Mircea Popa-Romania
1 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

www.ssmrmh.ro
Solutions by
Daniel Sitaru – Romania, Mehmet Şahin-Ankara-Turkyie,Marian Ursărescu-
Romania, D.M. Bătinețu-Giurgiu-Romania, Mohamed Amine Ben Ajiba-
Tanger-Morocco,Nguyen Van Canh-BenTre-Vietnam, Sanong Huayrerai-
Nakon Pathom-Thailand, Ravi Prakash-New Delhi-India, Daniel Văcaru-
Romania,George Apostolopoulos-Messolonghi-Greece, Ionuț Florin Voinea-
Romania,Florentin Vişescu-Romania, Agayev Seddredin-Azerbaijan, Jamal
Issah-Ghana, Hikmat Mammmadov-Azerbaijan, Khaled Abd Imouti-
Damascus-Syria,George Florin Şerban-Romania, Fayssal Abdelli-Bejaia-
Algerie, Avishek Mitra-West Bengal-India, Angel Plaza-Spain, Ertan Yildirim-
Turkiye, Cristian Miu-Romania, Soumitra Mandal-Chandar Nagore-India,
Amrit Awasthi-India, Neculai Stanciu-Romania, Florică Anastase-
Romania,Adrian Popa-Romania,Kamel Gandouli Rezgui-Tunisia, Ruxandra
Daniela Tonilă-Romania,Syed Shahabudeen-Kerala-India, Kaushik Mahanta-
Assam-India, Vasile Mircea Popa-Romania, Rana Ranino-Setif-Algerie,
Asmat Qatea-Afghanistan

www.ssmrmh.ro
PROBLEMS FOR JUNIORS
JP.391 In 𝚫𝑨𝑩𝑪, 𝑷 −inner point, 𝑴, 𝑳 ∈ [𝑨𝑩], 𝑫, 𝑬 ∈ [𝑩𝑪], 𝑭, 𝑲 ∈ [𝑪𝑨]
𝑨𝑴 = 𝑨𝑭, 𝑩𝑳 = 𝑩𝑬, 𝑪𝑲 = 𝑪𝑫, |𝑫𝑬| = 𝒂𝟏 , |𝑭𝑲| = 𝒃𝟏 , |𝑳𝑴| = 𝒄𝟏
(𝑴, 𝑷, 𝑭), (𝑪, 𝑷, 𝑳), (𝑫, 𝑷, 𝑲) −are collinear.Prove that:
𝟏
𝑭= (𝒂 𝒓 + 𝒃𝟏 𝒓𝒃 + 𝒄𝟏 𝒓𝒄 )
𝟐 𝟏 𝒂
Proposed by Mehmet Şahin-Ankara-Turkyie
Solution by proposer
̂
̂ ) = 𝟗𝟎∘ − 𝑨
̂ ) = 𝝁(𝑨𝑭𝑴
𝝁(𝑨𝑴𝑷 𝟐
̂
𝑩
̂ ) = 𝝁(𝑩𝑬𝑳
𝝁(𝑩𝑳𝑬 ̂ ) = 𝟗𝟎∘ −
𝟐
̂
𝑪
̂ ) = 𝟗𝟎∘ −
𝝁(𝑪𝑲𝑫
𝟐
Let 𝑷𝑻 ⊥ 𝑩𝑪, |𝑳𝑻| = 𝒉𝟏 .
𝒉𝟏 𝒂𝟏
𝚫𝑷𝑫𝑬~𝚫𝑰𝒂 𝑪𝑩 ⇒ =
𝒓𝒂 𝒂
𝒂𝟏
⇒ 𝒉𝟏 = ⋅ 𝒓𝒂
𝒂
𝒃𝟏 𝒄𝟏
Similarly, 𝒉𝟐 = ⋅ 𝒓𝒃 and 𝒉𝟑 = ⋅
𝒃 𝒄
𝒓𝒄 .
𝟏 𝟏 𝟏
𝑭 = [𝑷𝑩𝑪] + [𝑷𝑪𝑨] + [𝑷𝑨𝑩] = 𝒂 ⋅ 𝒉𝟏 + 𝒃 ⋅ 𝒉 𝟐 + 𝒄 ⋅ 𝒉𝟑 =
𝟐 𝟐 𝟐
𝟏 𝒂𝟏 𝟏 𝒃𝟏 𝟏 𝒄𝟏 𝟏
= ( ⋅ 𝒓𝒂 ) 𝒂 + ( ⋅ 𝒓𝒃 ) 𝒃 + ( ⋅ 𝒓𝒄 ) 𝒄 = (𝒂𝟏 𝒓𝒂 + 𝒃𝟏 𝒓𝒃 + 𝒄𝟏 𝒓𝒄 )
𝟐 𝒂 𝟐 𝒃 𝟐 𝒄 𝟐
JP.392 𝒛𝟏 , 𝒛𝟐 , 𝒛𝟑 ∈ ℂ∗ different in pairs such that
|𝒛𝟏 | = |𝒛𝟐 | = |𝒛𝟑 | = 𝟏, 𝑨(𝒛𝟏 ), 𝑩(𝒛𝟐 ), 𝑪(𝒛𝟑 ). Prove that:
𝒛𝟐 𝒛𝟑 𝟏
𝐈𝐟 ∑ = 𝐭𝐡𝐞𝐧 𝑨𝑩 = 𝑩𝑪 = 𝑪𝑨.
𝟏𝟒𝒛𝟐 𝒛𝟑 − 𝒛𝟐𝟐 − 𝒛𝟐𝟑 𝟓
𝒄𝒚𝒄
Proposed by Marian Ursărescu-Romania

www.ssmrmh.ro
Solution 1 by proposer
𝑨(𝒛𝟏 ), 𝑩(𝒛𝟐 ), 𝑪(𝒛𝟑 ), 𝚫𝑨𝑩𝑪 ⊂ 𝑪(𝟎, 𝟏)
𝟐
( 𝒛𝟐 − 𝒛𝟑 ) 𝟐 𝟒𝒛𝟐 𝒛𝟑 𝟒
𝐬𝐢𝐧 𝑨 = − ⇒∑ = ⇔
𝟒𝒛𝟐 𝒛𝟑 𝟏𝟐𝒛𝟐 𝒛𝟑 − (𝒛𝟐 − 𝒛𝟑 )𝟐 𝟓
𝒄𝒚𝒄
𝟏 𝟒 𝟏 𝟒
∑ = ⇔ ∑ = ; ( 𝟏)
( 𝒛𝟐 − 𝒛𝟑 ) 𝟐 𝟓 𝟑 + 𝐬𝐢𝐧𝟐 𝑨 𝟓
𝒄𝒚𝒄 𝟑 − 𝒄𝒚𝒄
𝟒𝒛𝟐 𝒛𝟑
𝟏 𝟒
𝐁𝐮𝐭: ∑ 𝟐
≥ ; (𝟐). 𝐅𝐫𝐨𝐦 (𝟏)&(𝟐) ⇒ 𝑨𝑩 = 𝑩𝑪 = 𝑪𝑨.
𝟑 + 𝐬𝐢𝐧 𝑨 𝟓
𝒄𝒚𝒄
𝟏 𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎 𝟗
∑ ≥
𝟑 + 𝐬𝐢𝐧𝟐 𝑨 𝟗 + ∑ 𝐬𝐢𝐧𝟐 𝑨
𝒄𝒚𝒄 𝟗 𝟒
𝟐 𝟐 ⇒ ≥ ⇔
𝒔 − 𝒓 − 𝟒𝑹𝒓 𝒔𝟐 𝟐
− 𝒓 − 𝟒𝑹𝒓 𝟓
∑ 𝐬𝐢𝐧𝟐 𝑨 = 𝟗+
𝟐𝑹𝟐 𝟐𝑹𝟐
{ 𝒄𝒚𝒄
𝟏𝟖𝑹𝟐 𝟒
𝟐 𝟐 𝟐
≥ ⇔ 𝟒𝟓𝑹𝟐 ≥ 𝟑𝟔𝑹𝟐 + 𝟐𝒔𝟐 − 𝟐𝒓𝟐 − 𝟖𝑹𝒓 ⇔
𝟏𝟖𝑹 + 𝒔 − 𝒓 − 𝟒𝑹𝒓 𝟓
𝟐𝒔𝟐 ≤ 𝟗𝑹𝟐 + 𝟖𝑹𝒓 + 𝟐𝒓𝟐
But: 𝒔𝟐 ≤ 𝟒𝑹𝟐 + 𝟒𝑹𝒓 + 𝟑𝒓𝟐 (𝑮𝒆𝒓𝒓𝒆𝒕𝒔𝒆𝒏) ⇒ 𝟐𝒔𝟐 ≤ 𝟖𝑹𝟐 + 𝟖𝑹𝒓 + 𝟔𝒓𝟐 ⇒
𝟖𝑹𝟐 + 𝟖𝑹𝒓 + 𝟔𝒓𝟐 ≤ 𝟒𝑹𝟐 + 𝟖𝑹𝒓 + 𝟐𝒓𝟐 ⇔ 𝟒𝒓𝟐 ≤ 𝑹𝟐 ⇔ 𝑹 ≥ 𝟐𝒓(𝑬𝒖𝒍𝒆𝒓).
Solution 2 by Mohamed Amine Ben Ajiba-Tanger-Morocco
( 𝒛 𝟐 − 𝒛𝟑 ) 𝟐
|𝒛𝟏 | = |𝒛𝟐 | = |𝒛𝟑 | = 𝟏 → 𝜟𝑨𝑩𝑪 ∈ 𝑪(𝑶, 𝟏) → 𝐬𝐢𝐧𝟐 𝑨 = − (𝑨𝒏𝒅 𝒂𝒏𝒂𝒍𝒐𝒈𝒔)
𝟒𝒛𝟐 𝒛𝟑
𝟏 𝟒 𝟏 𝟒
(∗) ↔ ∑ 𝟐
= ↔ ∑ 𝟐
=
( 𝒛𝟐 − 𝒛𝟑 ) 𝟓 𝟑 + 𝐬𝐢𝐧 𝑨 𝟓
𝒄𝒚𝒄 𝟑 − 𝒄𝒚𝒄
𝟒𝒛𝟐 𝒛𝟑
𝑪𝑩𝑺
𝟏 𝟗
𝑩𝒖𝒕 𝒘𝒆 𝒉𝒂𝒗𝒆 ∶ ∑ ⏞
≥
𝟐
𝟑 + 𝐬𝐢𝐧 𝑨 (𝟑 + 𝐬𝐢𝐧 𝑨) + (𝟑 + 𝐬𝐢𝐧𝟐 𝑩) + (𝟑 + 𝐬𝐢𝐧𝟐 𝑪)
𝟐
𝒄𝒚𝒄
𝟗
=
𝟗+ (𝐬𝐢𝐧𝟐 𝑨 + 𝐬𝐢𝐧𝟐 𝑩 + 𝐬𝐢𝐧𝟐 𝑪)
𝑳𝒆𝒊𝒃𝒏𝒊𝒛
𝒂𝟐 𝟏 𝟏 𝟗
𝑨𝒍𝒔𝒐, 𝒘𝒆 𝒉𝒂𝒗𝒆 ∶ ∑ 𝐬𝐢𝐧 𝑨 = ∑ 𝟐 = 𝟐
∑ 𝒂𝟐 ⏞
≤ . 𝟗𝑹𝟐 =
𝟒𝑹 𝟒𝑹𝟐 𝟒𝑹 𝟐 𝟒
𝟏 𝟒
→ ∑ ≥
𝟑 + 𝐬𝐢𝐧𝟐 𝑨 𝟓
𝒄𝒚𝒄
𝑾𝒊𝒕𝒉 𝒆𝒒𝒖𝒂𝒍𝒊𝒕𝒚 𝒊𝒇 𝜟𝑨𝑩𝑪 𝒊𝒔 𝒆𝒒𝒖𝒊𝒍𝒂𝒕𝒆𝒓𝒂𝒍.

www.ssmrmh.ro
𝒛𝟐 𝒛𝟑 𝟏
𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆, ∑ 𝟐 𝟐
= → 𝑨𝑩 = 𝑩𝑪 = 𝑪𝑨.
𝟏𝟒𝒛𝟐 𝒛𝟑 − 𝒛𝟐 − 𝒛𝟑 𝟓
𝒄𝒚𝒄
JP.393 In 𝚫𝑨𝑩𝑪 the following relationship holds:

𝟖𝟏
𝒉𝟑𝒂 + 𝒉𝟑𝒃 + 𝒉𝟑𝒄 ≤ (𝟗𝑹𝟑 − 𝟔𝟒𝒓𝟑 )
𝟖
Using inequality: (𝒙 + 𝒚 + 𝒛)𝟑 ≥ 𝒙𝟑 + 𝒚𝟑 + 𝒛𝟑 + 𝟐𝟒𝒙𝒚𝒛, ∀𝒙, 𝒚, 𝒛 > 0
(⇔ (𝒂 + 𝒃)(𝒃 + 𝒄)(𝒄 + 𝒂) ≥ 𝟖𝒂𝒃𝒄)
𝒉𝟑𝒂 + 𝒉𝟑𝒃 + 𝒉𝟑𝒄 ≤ (𝒉𝒂 + 𝒉𝒃 + 𝒉𝒄 )𝟑 − 𝟐𝟒𝒉𝒂 𝒉𝒃 𝒉𝒄 ; (𝟏)
𝒔𝟐 + 𝒓𝟐 + 𝟒𝑹𝒓 𝟐𝒔𝟐 𝒓𝟐
𝒉𝒂 + 𝒉𝒃 + 𝒉𝒄 = 𝐚𝐧𝐝 𝒉𝒂 𝒉𝒃 𝒉𝒄 = ; ( 𝟐)
𝟐𝑹 𝑹
From (𝟏), (𝟐) it follows that:
(𝒔𝟐 + 𝒓𝟐 + 𝟒𝑹𝒓)𝟑 𝟐𝒔𝟐 𝒓𝟐
𝒉𝟑𝒂 + 𝒉𝟑𝒃 + 𝒉𝟑𝒄 ≤ − 𝟐𝟒 ⋅ ; ( 𝟑)
𝟖𝑹𝟑 𝑹
But 𝒔𝟐 ≤ 𝟒𝑹𝟔𝟐 + 𝟒𝑹𝒓 + 𝟑𝒓𝟑 (𝑮𝒆𝒓𝒓𝒆𝒕𝒔𝒆𝒏) 𝐚𝐧𝐝 𝟐𝒔𝟐 ≥ 𝟐𝟕𝑹𝒓(𝑻𝒖𝒓𝒕𝒐𝒊𝒖); (𝟒)
From (𝟑), (𝟒) it follows that:
(𝟒𝑹𝟐 + 𝟖𝑹𝒓 + 𝟒𝒓𝟐 )𝟑 𝟐𝟒 ⋅ 𝟐 + 𝑹𝒓𝟑 𝟖(𝑹 + 𝒓)𝟔 𝑬𝒖𝒍𝒆𝒓
𝒉𝟑𝒂 + 𝒉𝟑𝒃 + 𝒉𝟑𝒄 ≤ − = − 𝟐𝟒 ⋅ 𝟐 + 𝑹 𝟑
≤
𝟖𝑹𝟑 𝑹 𝑹𝟑
𝟑𝑹 𝟔
( ) 𝟑𝟔 ⋅ 𝑹 𝟑 𝟗𝑹𝟑 − 𝟐𝟔 𝒓𝟑 𝟖𝟏
≤ 𝟖 ( 𝟐 𝟑 − 𝟖𝟏𝒓𝟑 ) = 𝟖 ( − 𝟖𝟏𝒓 𝟑
) = 𝟖 ⋅ 𝟖𝟏 ⋅ = (𝟗𝑹𝟑 − 𝟔𝟒𝒓𝟑 )
𝑹 𝟐𝟔 𝟐𝟔 𝟖

𝑾𝒆 𝒌𝒏𝒐𝒘 𝒕𝒉𝒂𝒕 ∶ 𝒉𝒂 ≤ 𝒘𝒂 ≤ √𝒔(𝒔 − 𝒂) = √𝒓𝒃 𝒓𝒄 (𝑨𝒏𝒅 𝒂𝒏𝒂𝒍𝒐𝒈𝒔)
𝑨𝑴−𝑮𝑴
𝒓𝒃 + 𝒓𝒄 𝟏
→ ∑ 𝒉𝒂 𝟑 ≤ ∑ 𝒓𝒃 𝒓𝒄 √𝒓𝒃 𝒓𝒄 ⏞
≤ ∑ 𝒓𝒃 𝒓𝒄 . = [∏(𝒓𝒂 + 𝒓𝒃 ) − 𝟐𝒓𝒂 𝒓𝒃 𝒓𝒄 ]
𝟐 𝟐
𝟏
= (𝟒𝑹𝒔𝟐 − 𝟐𝒓𝒔𝟐 ) =
𝟐
𝑴𝒊𝒕𝒓𝒊𝒏𝒐𝒗𝒊𝒄 ?
𝟐𝟕 𝟑 𝟖𝟏
= 𝟐𝑹𝒔 − 𝒓𝒔 𝟐 𝟐 ⏞
≤ 𝟑 ⏞
𝑹 − 𝟐𝟕𝒓 ≤ (𝟗𝑹𝟑 − 𝟔𝟒𝒓𝟑 )
𝟐 𝟖

www.ssmrmh.ro
↔ 𝟒𝟗𝟔𝟖𝒓 ≤ 𝟔𝟐𝟏𝑹 ↔ 𝟐𝒓)𝟑 ≤ 𝑹𝟑 ↔ 𝟐𝒓 ≤ 𝑹 (𝑬𝒖𝒍𝒆𝒓′ 𝒔 𝒊𝒏𝒆𝒒𝒖𝒂𝒍𝒊𝒕𝒚)
𝟑 ( 𝟑
𝟖𝟏
𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆, ∑ 𝒉𝒂 𝟑 ≤ (𝟗𝑹𝟑 − 𝟔𝟒𝒓𝟑 ).
𝟖
Solution 3 by Nguyen Van Canh-Ben Tre-Vietnam
𝒉𝒂 ≤𝒎𝒂 𝟐𝒃𝟐 + 𝟐𝒄𝟐 − 𝒂𝟐 𝟐𝑭
∑ 𝒉𝟑𝒂 = ∑(𝒉𝟐𝒂 ⋅ 𝒉𝒂 ) ≤ ∑(𝒎𝟐𝒂 ⋅ 𝒉𝒂 ) = ∑ ⋅ =
𝟒 𝒂
𝒄𝒚𝒄 𝒄𝒚𝒄 𝒄𝒖𝒄 𝒄𝒚𝒄
𝑭 𝟐(𝒃𝟐 + 𝒄𝟐 ) − 𝒂𝟐 𝑭 𝟐∑𝒃𝒄(𝒃𝟐 + 𝒄𝟐 ) − 𝒂𝒃𝒄∑𝒂 𝟐∑𝒃𝒄(𝒃𝟐 + 𝒄𝟐 ) − 𝒂𝒃𝒄∑𝒂

= ∑ = ⋅ = ≤
𝟐 𝒂 𝟐 𝒂𝒃𝒄 𝟖𝑹
𝒄𝒚𝒄
𝒃𝒄(𝒃𝟐 +𝒄𝟐)≤𝒃𝟒 +𝒄𝟒 𝟒(𝒂𝟒 + 𝒃𝟒 + 𝒄𝟒 ) − 𝒂𝒃𝒄∑𝒂 𝟖(𝒔𝟒 − (𝟗𝑹𝒓 + 𝟏𝟐𝒓𝟐 )𝒔𝟐 + (𝟒𝑹𝒓 + 𝒓𝟐 )𝟐 ) (𝟏)
≤ = ≤
𝟖𝑹 𝟖𝑹
𝟖𝟏(𝟗𝑹𝟑 − 𝟔𝟒𝒓𝟑 )
≤
𝟖
(𝟏) ⇔ 𝟖(𝒔𝟒 − (𝟗𝑹𝒓 + 𝟔𝒓𝟐 )𝒔𝟐 + (𝟒𝑹𝒓 + 𝒓𝟐 )𝟐 ) ≤ 𝟖𝟏(𝟗𝑹𝟒 − 𝟔𝟒𝑹𝒓𝟑 )
Let 𝒇(𝒖) = 𝒖𝟐 − (𝟗𝑹𝒓 + 𝟔𝒓𝟐 )𝒖 + (𝟒𝑹𝒓 + 𝒓𝟐 )𝟐 ; (∵ 𝒖 = 𝒔𝟐 )
𝒖≥𝟏𝟔𝑹𝒓−𝟓𝒓𝟐
𝒇′ (𝒖) = 𝟐𝒖 − (𝟗𝑹𝒓 + 𝟔𝒓𝟐 ) ≥ 𝟐(𝟏𝟔𝑹𝒓 − 𝟓𝒓𝟐 ) − (𝟗𝑹𝒓 + 𝟔𝒓𝟐 )
𝑹≥𝟐𝒓
= 𝟐𝟑𝑹𝒓 − 𝟏𝟔𝒓𝟐 ≥ 𝟑𝟎𝒓𝟐 ⇒ 𝒇′ (𝒖) ≥ 𝟎 ⇒ 𝒇(𝒖) ↗ [𝟏𝟔𝑹𝒓 − 𝟓𝒓𝟐 , 𝟒𝑹𝟐 + 𝟒𝑹𝒓 + 𝟑𝒓𝟐 ]
(𝟐)
⇒ 𝟖[(𝟒𝑹𝟐 + 𝟒𝑹𝒓 + 𝟑𝒓𝟐 )𝟐 − (𝟗𝑹𝒓 + 𝟔𝒓𝟐 )(𝟒𝑹𝟐 + 𝟒𝑹𝒓 + 𝟑𝒓𝟐 ) + (𝟒𝑹𝒓 + 𝒓𝟐 )𝟐 ] ≤
≤ 𝟖𝟏(𝟗𝑹𝟒 − 𝟔𝟒𝑹𝒓𝟑 )
(𝟐) ⇔ 𝟖[(𝟒𝒕𝟐 + 𝟒𝒕 + 𝟑) − (𝟗𝒕 + 𝟔)(𝟒𝒕𝟐 + 𝟒𝒕 + 𝟑) + (𝟒𝒕 + 𝟏)𝟐 ]
𝑹
≤ 𝟖𝟏(𝟗𝒕𝟒 − 𝟔𝟒𝒕), (∵ 𝒕 = ≥ 𝟐)
𝒓
⇔ 𝟖(𝟏𝟔𝒕𝟒 − 𝟒𝒕𝟑 − 𝟒𝒕𝟐 − 𝟏𝟗𝒕 − 𝟖) ≤ 𝟕𝟐𝟗𝒕𝟒 − 𝟓𝟏𝟖𝟒𝒕
⇔ 𝟔𝟎𝟏𝒕𝟒 + 𝟑𝟐𝒕𝟑 + 𝟑𝟐𝒕𝟐 − 𝟓𝟎𝟑𝟐𝒕 + 𝟔𝟒 ≥ 𝟎
⇔ (𝒕 − 𝟐)(𝟔𝟎𝟏𝒕𝟑 + 𝟏𝟐𝟑𝟒𝒕𝟐 + 𝟐𝟓𝟎𝟎𝒕 − 𝟑𝟐) ≥ 𝟎
Which is true from 𝒕 ≥ 𝟐 ⇒ 𝒕 − 𝟐 ≥ 𝟎 and
𝒕≥𝟐
𝟔𝟎𝟏𝒕𝟑 + 𝟏𝟐𝟑𝟒𝒕𝟐 + 𝟐𝟓𝟎𝟎𝒕 − 𝟑𝟐 > 2500𝑡 − 32 > 16𝒕 − 𝟑𝟐 > 𝟏𝟔 ⋅ 𝟐 − 𝟑𝟐 = 𝟎
⇒ (𝟐) ⇒ (𝟏) it’s true.

www.ssmrmh.ro
JP.394 If 𝒙, 𝒚, 𝒛, 𝒕 > 0 then:
𝒙𝟒 + 𝟏 𝒚𝟒 + 𝟏 𝒛𝟒 + 𝟏 𝒕𝟒 + 𝟏
+ + + ≥𝟖
𝒚√𝒙𝟒 − 𝒙𝟐 + 𝟏 𝒛√𝒚𝟒 − 𝒚𝟐 + 𝟏 𝒛√𝒛𝟒 − 𝒛𝟐 + 𝟏 𝒙√𝒕𝟒 − 𝒕𝟐 + 𝟏
Proposed by Daniel Sitaru-Romania
𝒙𝟒 + 𝟏 𝒙𝟒 − 𝒙𝟐 + 𝟏 + 𝒙𝟐
∑ =∑ =
𝒄𝒚𝒄
𝒚√𝒙𝟒 − 𝒙𝟐 + 𝟏 𝒄𝒚𝒄
𝒚√𝒙𝟒 − 𝒙𝟐 + 𝟏
𝟏 𝒙𝟒 − 𝒙𝟐 + 𝟏 𝒙𝟐 𝟏 𝒙𝟐 𝑨𝑴−𝑮𝑴
=∑ ( + ) = ∑ (√𝒙𝟒 − 𝒙𝟐 + 𝟏 + ) ≥
𝒚 √𝒙𝟒 − 𝒙𝟐 + 𝟏 √𝒙𝟒 − 𝒙𝟐 + 𝟏 𝒚 √𝒙𝟒 − 𝒙𝟐 + 𝟏
𝒄𝒚𝒄 𝒄𝒚𝒄
𝟏 𝒙𝟐 𝒙 𝑨𝑴−𝑮𝑴 𝟒 𝒙 𝒚 𝒛 𝒕
≥ ∑ ⋅ 𝟐√√𝒙𝟒 − 𝒙𝟐 + 𝟏 ⋅ = 𝟐∑ ≥ 𝟐⋅𝟒⋅ √ ⋅ ⋅ ⋅ =𝟖
𝒚 √𝒙𝟒 − 𝒙𝟐 + 𝟏 𝒚 𝒚 𝒛 𝒕 𝒙
Equality holds for 𝒙 = 𝒚 = 𝒛 = 𝒕 = 𝟏.

Solution 2 by Sanong Huayrerai-Nakon Pathom-Thailand
𝒙𝟒 + 𝟏
∑ ≥
𝒄𝒚𝒄
𝒚√𝒙𝟒 − 𝒙𝟐 + 𝟏
𝟒 (𝒙𝟒 + 𝟏)(𝒚𝟒 + 𝟏)(𝒛𝟒 + 𝟏)(𝒕𝟒 + 𝟏)

≥ 𝟒⋅ √ ≥𝟖
𝒙𝒚𝒛𝒕√(𝒙𝟒 − 𝒙𝟐 + 𝟏)(𝒚𝟒 − 𝒚𝟐 + 𝟏)(𝒛𝟒 − 𝒛𝟐 + 𝟏)(𝒕𝟒 − 𝒕𝟐 + 𝟏)
[(𝒙𝟒 + 𝟏)(𝒚𝟒 + 𝟏)(𝒛𝟒 + 𝟏)(𝒕𝟒 + 𝟏)]𝟐 ≥

≥ 𝟒(𝒙𝒚𝒛𝒕)𝟐 (𝒙𝟒 − 𝒙𝟐 + 𝟏)(𝒚𝟒 − 𝒚𝟐 + 𝟏)(𝒛𝟒 − 𝒛𝟐 + 𝟏)(𝒕𝟒 − 𝒕𝟐 + 𝟏)
Because (𝒙𝟒 + 𝟏)𝟐 ≥ 𝟒𝒙𝟐 (𝒙𝟒 − 𝒙𝟐 + 𝟏) = 𝟒𝒙𝟔 − 𝟒𝒙𝟒 + 𝟒𝒙𝟐
𝒙𝟖 + 𝟐𝒙𝟒 + 𝟒𝒙𝟒 ≥ 𝟒𝒙𝟔 + 𝟒𝒙𝟐
( 𝒙 𝟖 − 𝒙 𝟔 ) + 𝟑( 𝒙 𝟒 − 𝒙 𝟐 ) ≥ 𝟑( 𝒙 𝟔 − 𝒙 𝟒 ) + ( 𝒙 𝟐 − 𝟏)
𝒙𝟔 (𝒙𝟐 − 𝟏) + 𝟑𝒙𝟐 (𝒙𝟐 − 𝟏) ≥ 𝟑𝒙𝟒 (𝒙𝟐 − 𝟏) + (𝒙𝟐 − 𝟏) = (𝒙𝟐 − 𝟏)(𝟑𝒙𝟒 + 𝟏)
(𝒙𝟐 − 𝟏) (𝒙𝟔 + 𝟑𝒙𝟐 − (𝟑𝒙𝟒 + 𝟏)) ≥ 𝟎, which is true because
for 𝒙 ≥ 𝟏: 𝒙𝟔 + 𝟑𝒙𝟐 ≥ 𝟑𝒙𝟒 + 𝟏 and for 𝒙 < 1: 𝒙𝟔 + 𝟑𝒙𝟐 < 3𝒙𝟒 + 𝟏

Analogous,
(𝒚𝟒 + 𝟏)𝟐 ≥ 𝟒𝒚𝟐 (𝒚𝟒 − 𝒚𝟐 + 𝟏), (𝒛𝟒 + 𝟏)𝟐 ≥ 𝟒𝒛𝟐 (𝒛𝟒 − 𝒛𝟐 + 𝟏),
www.ssmrmh.ro
and (𝒕𝟒 + 𝟏)𝟐 ≥ 𝟒𝒕𝟐 (𝒕𝟒 − 𝒕𝟐 + 𝟏)
Solution 3 by Ravi Prakash-New Delhi-India

We first show that for 𝒙 > 0:
𝟏
𝒙𝟒 + 𝟏 𝟐 𝒙𝟐 +
≥𝟐⇔ 𝒙 ≥𝟐⇔
𝒙√𝒙𝟒 − 𝒙𝟐 + 𝟏 𝟏
√ 𝒙𝟐 + 𝟐 − 𝟐
𝒙
𝟏 𝟏
√ 𝒙𝟐 + −𝟏+ ≥ 𝟐, 𝐭𝐫𝐮𝐞 𝐟𝐨𝐫 (𝑨𝑴 − 𝑮𝑴)
𝒙𝟐 𝟏
√ 𝒙𝟐 + 𝟐−𝟏
𝒙
Now,
𝒙𝟒 + 𝟏 𝟒 𝒙𝟒 + 𝟏 𝟒
∑ ≥ 𝟒 ⋅ √∏ ≥ 𝟒 ⋅ √ 𝟐𝟒 = 𝟖
𝒄𝒚𝒄
𝒚√𝒙𝟒 − 𝒙𝟐 +𝟏 𝒄𝒚𝒄
𝒙√𝒙𝟒 − 𝒙𝟐 +𝟏

𝒙𝟒 + 𝟏
𝐖𝐞 𝐡𝐚𝐯𝐞 ∶
√𝒙𝟒 − 𝒙𝟐 + 𝟏
𝑨𝑴−𝑮𝑴
𝒙𝟐 𝒙𝟐
= √𝒙𝟒 − 𝒙𝟐 + 𝟏 + ⏞
≥ 𝟐√√𝒙𝟒 − 𝒙𝟐 + 𝟏. = 𝟐𝒙
√𝒙𝟒 − 𝒙𝟐 + 𝟏 √𝒙𝟒 − 𝒙𝟐 + 𝟏
𝒙𝟒 + 𝟏 𝒙
→ ≥ 𝟐.
𝒚√𝒙𝟒 − 𝒙𝟐 + 𝟏 𝒚
𝐒𝐢𝐦𝐢𝐥𝐚𝐫𝐥𝐲, 𝐰𝐞 𝐡𝐚𝐯𝐞 ∶
𝒚𝟒 + 𝟏 𝒚 𝒛𝟒 + 𝟏 𝒛 𝒕𝟒 + 𝟏 𝒕
≥ 𝟐. , ≥ 𝟐. 𝒂𝒏𝒅 ≥ 𝟐.
𝒛√𝒚𝟒 − 𝒚𝟐 + 𝟏 𝒛 𝒕√𝒛𝟒 − 𝒛𝟐 + 𝟏 𝒕 𝒙√𝒕𝟒 − 𝒕𝟐 + 𝟏 𝒙
𝑨𝑴−𝑮𝑴
𝒙 𝒚 𝒛 𝒕 𝟒 𝒙 𝒚 𝒛 𝒕
→ 𝑳𝑯𝑺(∗) ≥ 𝟐( + + + ) ⏞
≥ 𝟐. 𝟒√ . . . = 𝟖.
𝒚 𝒛 𝒕 𝒙 𝒚 𝒛 𝒕 𝒙
Solution 5 by Daniel Văcaru-Romania

We have:
𝒙𝟒 + 𝟏 𝒙𝟓 + 𝒙 𝒙𝟓 + 𝒙 𝑨𝑴−𝑮𝑴
= = ≥
𝒚√𝒙𝟒 − 𝒙𝟐 + 𝟏 𝒚𝒙√𝒙𝟒 − 𝒙𝟐 + 𝟏 𝒚√𝒙𝟐 (𝒙𝟒 − 𝒙𝟐 + 𝟏)

www.ssmrmh.ro
𝟐(𝒙𝟓 + 𝒙) 𝟐𝒙
≥ 𝟐 𝟒 𝟐
=
𝒚[𝒙 + (𝒙 − 𝒙 + 𝟏)] 𝒚
Therefore,
𝑨𝑴−𝑮𝑴
𝒙𝟒 + 𝟏 𝒙 𝒚 𝒛 𝒕 𝟒 𝒙 𝒚 𝒛 𝒕
∑ ≥ 𝟐( + + + ) ⏞
≥ 𝟐. 𝟒√ . . . = 𝟖.
𝒚√𝒙𝟒 − 𝒙𝟐 + 𝟏 𝒚 𝒛 𝒕 𝒙 𝒚 𝒛 𝒕 𝒙
𝒄𝒚𝒄
JP.395 Let 𝑹 and 𝒓 be the circumradius and inradius, respectively, of triangle

𝑨𝑩𝑪. Let 𝑫, 𝑬 and 𝑭 be chosen on sides 𝑩𝑪, 𝑪𝑨 and 𝑨𝑩, so that 𝑨𝑫, 𝑩𝑬 and
𝑫𝑬 𝑬𝑭 𝑭𝑫 𝟑 𝑹
𝑪𝑭 bisect the angles of 𝑨𝑩𝑪. Prove + + ≤ (𝟏 + ).
𝑨𝑩 𝑩𝑪 𝑪𝑨 𝟒 𝟐𝒓
Proposed by George Apostolopoulos-Messolonghi-Greece

𝒃𝒄
Let 𝒂 = 𝑩𝑪, 𝒃 = 𝑪𝑨, 𝒄 = 𝑨𝑩 be the lengths of the sides. We know that 𝑨𝑭 = 𝒂+𝒃,
𝒃𝒄
𝑨𝑬 = .
𝒂+𝒄
Using the law of cosines in triangle 𝑨𝑭𝑬, we have 𝑬𝑭𝟐 = 𝑨𝑭𝟐 + 𝑨𝑬𝟐 − 𝟐𝑨𝑭 ⋅ 𝑨𝑬 ⋅ 𝐜𝐨𝐬 𝑨.
𝒃𝟐 +𝒄𝟐−𝒂𝟐
Also, we know that: 𝐜𝐨𝐬 𝑨 = . So,
𝟐𝒃𝒄
𝟐 𝟐 𝟐 𝟐
𝟐
𝒃𝒄 𝒃𝒄 𝟐𝒃 𝒄 𝒃𝟐 + 𝒄𝟐 − 𝒂𝟐
𝑬𝑭 = ( ) +( ) − ⋅ =
𝒂+𝒃 𝒂+𝒄 (𝒂 + 𝒃)(𝒂 + 𝒄) 𝟐𝒃𝒄
𝟏 𝟏 𝒃𝟐 + 𝒄𝟐 − 𝒂𝟐
= 𝒃𝟐 𝒄𝟐 ( + − )=
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐 𝒃𝒄(𝒂 + 𝒃)(𝒂 + 𝒄)
𝟐 𝟐
𝒃𝒄(𝒂 + 𝒄)𝟐 + 𝒃𝒄(𝒂 + 𝒃)𝟐 − (𝒃𝟐 + 𝒄𝟐 − 𝒂𝟐 )(𝒂 + 𝒃)(𝒂 + 𝒄)
=𝒃 𝒄 ⋅ =
𝒃𝒄(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐
𝒃𝟐 𝒄𝟐
= ⋅
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐

www.ssmrmh.ro
𝒃𝒄(𝒃𝟐 + 𝒄𝟐 ) + 𝟐𝒂𝒃𝒄(𝒃 + 𝒄) + 𝟐𝒂𝟐 𝒃𝒄 − 𝒂(𝒂 + 𝒃 + 𝒄)(𝒃𝟐 + 𝒄𝟐 − 𝒂𝟐 ) − 𝒃𝒄(𝒃𝟐 + 𝒄𝟐 ) + 𝒂𝟐 𝒃𝒄
⋅
𝒃𝒄
𝒃𝟐 𝒄𝟐
= ⋅
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐
𝒂𝟐 (𝒃𝒄 + 𝒂𝟐 + 𝒂𝒃 + 𝒂𝒄) + 𝟐𝒂𝒃𝒄(𝒂 + 𝒃 + 𝒄) − 𝒂(𝒂 + 𝒃 + 𝒄)(𝒃𝟐 + 𝒄𝟐 )
⋅
𝒃𝒄
𝟐
𝒃𝒄 𝒂𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄) − 𝒂(𝒂 + 𝒃 + 𝒄)(𝒃𝟐 + 𝒄𝟐 − 𝟐𝒃𝒄)
=( ) ⋅ =
(𝒂 + 𝒃)(𝒂 + 𝒄) 𝒃𝒄
𝟐
𝒃𝒄 𝒂𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄) − 𝒂(𝒂 + 𝒃 + 𝒄)(𝒃 − 𝒄)𝟐
=( ) ⋅ ≤
(𝒂 + 𝒃)(𝒂 + 𝒄) 𝒃𝒄
𝟐
𝒃𝒄 𝒂𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄) 𝒂𝟐 𝒃𝒄
≤( ) ⋅ = .
(𝒂 + 𝒃)(𝒂 + 𝒄) 𝒃𝒄 (𝒂 + 𝒃)(𝒂 + 𝒄)
𝟐
𝒂 + √𝒃𝒄
𝒂𝟐 𝒃𝒄 𝒂𝟐 𝒃𝒄 𝒂√𝒃𝒄 √( 𝟐 )
𝑬𝑭 ≤ √ ≤√ =√ = =
(𝒂 + 𝒃)(𝒂 + 𝒄) 𝟐√𝒂𝒃 ⋅ 𝟐√𝒂𝒄 𝟒 𝟒
𝒃+𝒄
𝒂+√𝒃𝒄 𝒂+ 𝟐𝒂+𝒃+𝒄 𝟐𝒂+𝒃+𝒄
𝟐
= ≤
= 𝟖 . Namely, 𝑬𝑭 ≤ 𝟖 . Similarly,
𝟒 𝟒
𝒂 + 𝟐𝒃 + 𝒄 𝒂 + 𝒃 + 𝟐𝒄
𝑭𝑫 ≤ , 𝑫𝑬 ≤ .
𝟖 𝟖
Now, we have:
𝑫𝑬 𝑬𝑭 𝑭𝑫 𝟏 𝒂 + 𝒃 + 𝟐𝒄 𝟐𝒂 + 𝒃 + 𝒄 𝒂 + 𝟐𝒃 + 𝒄
+ + ≤ ( + + )=
𝑨𝑩 𝑩𝑪 𝑪𝑨 𝟖 𝒄 𝒂 𝒃
𝟏 𝒂 𝒃 𝒃 𝒄 𝒄 𝒂
= (𝟔 + ( + ) + ( + ) + ( + )) ; (∗)
𝟖 𝒃 𝒂 𝒄 𝒃 𝒂 𝒄
𝒂 𝒃 𝑹
Now, we’ll prove that + ≤ . Consider the substitutions 𝒂 = 𝒚 + 𝒛, 𝒃 = 𝒛 + 𝒙, 𝒄 = 𝒙 +
𝒃 𝒂 𝒓
𝑹 𝒂𝒃𝒄 𝒂+𝒃+𝒄
𝒚, where 𝒙, 𝒚, 𝒛 are positive real numbers. We know that = 𝟒(𝒔−𝒂)(𝒔−𝒃)(𝒔−𝒄) , 𝒔 = .
𝒓 𝟐
We have:
𝟏 𝟏 𝟏 𝟏 𝒙+𝒚
(𝒛+𝒙)𝟐
+ (𝒚+𝒛)𝟐 ≤ 𝟒𝒛𝒙 + 𝟒𝒙𝒚 = 𝟒𝒙𝒚𝒛 and multiply by (𝒛 + 𝒙)(𝒚 + 𝒛) both sides, we get
𝒚 + 𝒛 𝒛 + 𝒙 (𝒙 + 𝒚)(𝒚 + 𝒛)(𝒛 + 𝒙)
+ ≤
𝒛+𝒙 𝒚+𝒛 𝟒𝒙𝒚𝒛
𝒂 𝒃 𝑹 𝒃 𝒄 𝑹 𝒄 𝒂 𝑹
Namely, 𝒃 + 𝒂 ≤ 𝒓 . Similarly, 𝒄 + 𝒃 ≤ 𝒓 , 𝒂 + 𝒄 ≤ 𝒓 . Now, (∗) gives
𝑫𝑬 𝑬𝑭 𝑭𝑫 𝟏 𝑹 𝑹 𝑹 𝟑 𝑹
+ + ≤ (𝟔 + + + ) = (𝟏 + )
𝑨𝑩 𝑩𝑪 𝑪𝑨 𝟖 𝒓 𝒓 𝒓 𝟒 𝟐𝒓
So, 𝑨𝑩 + 𝑩𝑪 + 𝑪𝑨 ≤ 𝟒 (𝟏 + 𝟐𝒓).
Equality holds when the triangle 𝑨𝑩𝑪 is equilateral.
𝑳𝒆𝒕 𝒂 = 𝑩𝑪, 𝒃 = 𝑪𝑨, 𝒄 = 𝑨𝑩.

www.ssmrmh.ro
𝒃𝒄 𝒃𝒄
𝑾𝒆 𝒌𝒏𝒐𝒘 𝒕𝒉𝒂𝒕 ∶ 𝑨𝑬 = 𝒂𝒏𝒅 𝑨𝑭 = .
𝒂+𝒄 𝒂+𝒃
𝑼𝒔𝒊𝒏𝒈 𝒕𝒉𝒆 𝑳𝒂𝒘 𝒐𝒇 𝒄𝒐𝒔𝒊𝒏𝒆𝒔 𝒊𝒏 ∆𝑨𝑬𝑭 ∶ 𝑬𝑭𝟐 = 𝑨𝑬𝟐 + 𝑨𝑭𝟐 − 𝟐. 𝑨𝑬. 𝑨𝑭. 𝐜𝐨𝐬 𝑨 =
𝒃𝟐 𝒄𝟐 𝒃𝟐 𝒄𝟐 𝒃𝒄(𝒃𝟐 + 𝒄𝟐 − 𝒂𝟐 )
= + −
(𝒂 + 𝒄)𝟐 (𝒂 + 𝒃)𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄)
𝟐 𝟐
𝟏 𝟏 𝟐 𝒃𝒄(𝒃 − 𝒄)𝟐 𝒂𝟐 𝒃𝒄
=𝒃 𝒄 ( − ) − + =
𝒂+𝒄 𝒂+𝒃 (𝒂 + 𝒃)(𝒂 + 𝒄) (𝒂 + 𝒃)(𝒂 + 𝒄)
𝒃𝒄(𝒃 − 𝒄)𝟐 𝒂𝟐 𝒃𝒄
=− [( 𝒂 + 𝒃)( 𝒂 + 𝒄) − 𝒃𝒄] +
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄)
𝟐𝒔𝒂𝒃𝒄(𝒃 − 𝒄)𝟐 𝒂𝟐 𝒃𝒄
=− + ≤
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄)
𝑨𝑴−𝑮𝑴
𝒂𝟐 𝒃𝒄 𝑬𝑭 𝒃 𝒄 𝟏 𝒃 𝒄
≤ → ≤√ . ⏞
≤ ( + )
(𝒂 + 𝒃)(𝒂 + 𝒄) 𝑩𝑪 𝒂+𝒃 𝒂+𝒄 𝟐 𝒂+𝒃 𝒂+𝒄
𝑬𝑭 𝟏 𝒃 𝒄 𝟏 𝒃 𝒂 𝟏
→ ∑ ≤∑ ( + ) = ∑( + ) = ∑𝟏
𝑩𝑪 𝟐 𝒂+𝒃 𝒂+𝒄 𝟐 𝒂+𝒃 𝒂+𝒃 𝟐
𝑬𝒖𝒍𝒆𝒓
𝟑 𝟑 𝑹
⏞
= ( 𝟏 + 𝟏) ≤ (𝟏 + ).
𝟒 𝟒 𝟐𝒓
𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆, + + ≤ (𝟏 + ).
𝑨𝑩 𝑩𝑪 𝑪𝑨 𝟒 𝟐𝒓
JP.396 Let 𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 be the altitudes from the vertices 𝑨, 𝑩, 𝑪 respectively, 𝑹
the circumradius and 𝒓 the inradius of a triangle 𝑨𝑩𝑪. Let 𝑨𝟏 , 𝑩𝟏 and 𝑪𝟏 be
chose on the sides 𝑩𝑪, 𝑪𝑨 and 𝑨𝑩 so that 𝑨𝑨𝟏 , 𝑩𝑩𝟏 and 𝑪𝑪𝟏 bisect the
angles of 𝑨𝑩𝑪. Let 𝒉𝑨 , 𝒉𝒃 and 𝒉𝑪 denote the altitudes of triangles
𝑨𝑩𝟏 𝑪𝟏 , 𝑩𝑪𝟏 𝑨𝟏 and 𝑪𝑨𝟏 𝑩𝟏 from the vertices 𝑨, 𝑩, 𝑪, respectively.
𝟑 𝒉 𝟑 𝒉 𝟑 𝒉 𝟑 𝑹
Prove that: √ 𝒂 + √ 𝒃 + √ 𝒄 ≤ 𝟑 √𝟐 ⋅ .
𝒉 𝒉 𝑨 𝒉 𝑩 𝟐𝒓 𝑪

𝒃𝒄
Let 𝒂 = 𝑩𝑪, 𝒃 = 𝑪𝑨, 𝒄 = 𝑨𝑩 be the lengths of the sides. We know that 𝑨𝑪𝟏 = 𝒂+𝒃,
𝒃𝒄
𝑨𝑩𝟏 = 𝒂+𝒄. Using the law of the cosines in triangle 𝑨𝑩𝟏 𝑪𝟏 , we have
www.ssmrmh.ro
𝑩𝟏 𝑪𝟐𝟏
= 𝑨𝑪𝟐𝟏
+ 𝑨𝑩𝟐𝟏 − 𝟐𝑨𝑪𝟏 ⋅ 𝑨𝑩𝟏 ⋅ 𝐜𝐨𝐬 𝑨 =
𝟐
𝒃𝒄 𝒃𝒄 𝟐 𝟐𝒃𝟐 𝒄𝟐
=( ) +( ) − 𝐜𝐨𝐬 𝑨
𝒂+𝒃 𝒂+𝒄 (𝒂 + 𝒃)(𝒂 + 𝒄)
𝒃𝟐 +𝒄𝟐 −𝒂𝟐
We know that:𝐜𝐨𝐬 𝑨 = ; (𝒂𝒏𝒅 𝒂𝒏𝒂𝒍𝒐𝒈𝒔). So,
𝟐𝒃𝒄
𝒃𝒄 𝟐 𝒃𝒄 𝟐 𝟐𝒃𝟐 𝒄𝟐 𝒃𝟐 + 𝒄𝟐 − 𝒂𝟐
𝑩𝟏 𝑪𝟐𝟏=( ) +( ) − ⋅ =
𝒂+𝒃 𝒂+𝒄 (𝒂 + 𝒃)(𝒂 + 𝒄) 𝟐𝒃𝒄
𝟏 𝟏 𝒃𝟐 + 𝒄𝟐 − 𝒂𝟐
= 𝒃𝟐 𝒄𝟐 ( + − )=
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐 𝒃𝒄(𝒂 + 𝒃)(𝒂 + 𝒄)
𝟐 𝟐
𝒃𝒄(𝒂 + 𝒄)𝟐 + 𝒃𝒄(𝒂 + 𝒃)𝟐 − (𝒃𝟐 + 𝒄𝟐 − 𝒂𝟐 )(𝒂 + 𝒃)(𝒂 + 𝒄)
=𝒃 𝒄 ⋅ =
𝒃𝒄(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐
𝒃𝟐 𝒄𝟐 𝒂𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄) − 𝒂(𝒂 + 𝒃 + 𝒄)(𝒃𝟐 + 𝒄𝟐 − 𝟐𝒃𝒄)
= ⋅ =
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐 𝒃𝒄
𝒃𝟐 𝒄𝟐 𝒂𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄) − 𝒂(𝒂 + 𝒃 + 𝒄)(𝒃 − 𝒄)𝟐
= ⋅ ≤
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐 𝒃𝒄
𝒃𝟐 𝒄𝟐 𝒂𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄) 𝒂𝟐 𝒃𝒄
≤ ⋅ =
(𝒂 + 𝒃)𝟐 (𝒃 + 𝒄)𝟐 𝒃𝒄 (𝒂 + 𝒃)(𝒂 + 𝒄)
So,
𝟐
𝒂 + √𝒃𝒄
𝒂𝟐 𝒃𝒄 𝒂𝟐 𝒃𝒄 𝒂√𝒃𝒄 √( 𝟐 )
𝒂 + √𝒃𝒄
𝑩𝟏 𝑪𝟏 ≤ √ ≤ √ =√ ≤ = ≤
(𝒂 + 𝒃)(𝒂 + 𝒄) 𝟐√𝒂𝒃 ⋅ 𝟐√𝒂𝒄 𝟒 𝟒 𝟒
𝒃+𝒄
𝒂+
≤ 𝟐 = 𝟐𝒂 + 𝒃 + 𝒄.
𝟒 𝟖
𝟐𝒂+𝒃+𝒄 𝒂+𝟐𝒃+𝒄 𝒂+𝒃+𝟐𝒄
Namely, 𝑩𝟏 𝑪𝟏 ≤ 𝟖 . Similarly, 𝑨𝟏 𝑪𝟏 ≤ 𝟖 and 𝑨𝟏 𝑩𝟏 ≤ 𝟖 .
𝟏 𝟏
Now, in triangle 𝑨𝑩𝟏 𝑪𝟏 , we have:𝟐 𝑩𝟏 𝑪𝟏 ⋅ 𝒉𝑨 = 𝟐 𝑨𝑪𝟏 ⋅ 𝑨𝑩𝟏 ⋅ 𝐬𝐢𝐧 𝑨 or
𝒃𝒄 𝒃𝒄 𝒂
⋅ 𝒂 + 𝒄 ⋅ 𝟐𝑹 𝟏 𝟐𝑹(𝒂 + 𝒃)(𝒃 + 𝒄)𝑩𝟏 𝑪𝟏
𝒉𝑨 = 𝒂 + 𝒃 ⇔ =
𝑩𝟏 𝑪𝟏 𝒉𝑨 𝒂𝒃𝟐 𝒄𝟐
𝟐𝒂 + 𝒃 + 𝒄
𝟏 𝟐𝑹(𝒂 + 𝒃)(𝒃 + 𝒄)
⇒ ≤ 𝟖
𝒉𝑨 𝒂𝒃𝟐 𝒄𝟐
Also, we have: 𝒃𝒄 = 𝟐𝑹 ⋅ 𝒉𝒂. So,
𝟐𝒂 + 𝒃 + 𝒄 𝒂 + 𝒃 + 𝒂 + 𝒄 𝟐 𝟐𝒂 + 𝒃 + 𝒄 𝒃𝒄
𝒉𝒂 𝟐𝑹 ( 𝒂 + 𝒃)( 𝒂 + 𝒄) ⋅ ⋅ 𝒉 𝒂 𝟐𝑹 ( ) ⋅ ⋅ 𝟐𝑹
≤ 𝟖 ≤ 𝟐 𝟖
𝒉𝑨 𝒂𝒃𝟐 𝒄𝟐 𝒂𝒃𝟐 𝒄𝟐
(𝟐𝒂 + 𝒃 + 𝒄)𝟑
=
𝟑𝟐𝒂𝒃𝒄
𝒉 (𝟐𝒂+𝒃+𝒄)𝟑 𝟑 𝒉 𝟐𝒂+𝒃+𝒄 𝟑 𝒉 𝒂+𝟐𝒃+𝒄 𝟑 𝒉𝒄 𝒂+𝒃+𝟐𝒄
Namely, 𝒉𝒂 ≤ ⇔ √𝒉𝒂 ≤ 𝟑 . Similarly, √ 𝒉𝒃 ≤ 𝟑 , √𝒉 ≤ 𝟑 .
𝑨 𝟑𝟐𝒂𝒃𝒄 𝑨 √𝟑𝟐𝒂𝒃𝒄 𝑩 √𝟑𝟐𝒂𝒃𝒄 𝑪 √𝟑𝟐𝒂𝒃𝒄

www.ssmrmh.ro
Adding up these inequalities, we have:
𝟑 𝒉𝒂 𝟑 𝒉𝒃 𝟑 𝒉𝒄 𝟐(𝒂 + 𝒃 + 𝒄) 𝟒𝒔
√ +√ +√ ≤ 𝟑 =𝟑 𝟑
𝒉𝑨 𝒉𝑩 𝒉𝑪 √𝟒 ⋅ √𝟑 ⋅ 𝒓 √𝟒 ⋅ √𝒂𝒃𝒄
But 𝑹 ≥ 𝟐𝒓, 𝟐𝒔 ≤ 𝟑√𝟑𝑹, 𝒔 ≥ 𝟑√𝟑𝒓and 𝒂𝒃𝒄 = 𝟒𝑹𝒓𝒔, so
𝒉𝒂 𝟑 𝒉𝒃 𝟑 𝒉𝒄
𝟑 𝟔√𝟑𝑹 𝟑 𝑹
√ +√ +√ ≤𝟑 𝟑 = 𝟑√𝟐 ⋅
𝒉𝑨 𝒉𝑩 𝒉𝑪 √𝟒 ⋅ √𝟒𝑹𝒓𝒔 𝟐𝒓
Equality holds if and only if triangle 𝑨𝑩𝑪 is equilateral.

𝑳𝒆𝒕 𝒂 = 𝑩𝑪, 𝒃 = 𝑪𝑨, 𝒄 = 𝑨𝑩 𝒂𝒏𝒅 𝑭 𝒃𝒆 𝒕𝒉𝒆 𝒂𝒓𝒆𝒂 𝒐𝒇 ∆𝑨𝑩𝑪.
𝒃𝒄 𝒃𝒄
𝑾𝒆 𝒌𝒏𝒐𝒘 𝒕𝒉𝒂𝒕 ∶ 𝑨𝑩𝟏 = 𝒂𝒏𝒅 𝑨𝑪𝟏 = .
𝒂+𝒄 𝒂+𝒃
𝑼𝒔𝒊𝒏𝒈 𝒕𝒉𝒆 𝑳𝒂𝒘 𝒐𝒇 𝒄𝒐𝒔𝒊𝒏𝒆𝒔 𝒊𝒏 ∆𝑨𝑩𝟏 𝑪𝟏 ∶ 𝑩𝟏 𝑪𝟏 𝟐 = 𝑨𝑩𝟏 𝟐 + 𝑨𝑪𝟏 𝟐 − 𝟐. 𝑨𝑩𝟏 . 𝑨𝑪𝟏 . 𝐜𝐨𝐬 𝑨
=
𝟐 𝟐
𝒃 𝒄 𝒃𝟐 𝒄𝟐 𝒃𝒄(𝒃𝟐 + 𝒄𝟐 − 𝒂𝟐 )
= + −
(𝒂 + 𝒄)𝟐 (𝒂 + 𝒃)𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄)
𝟐 𝟐
𝒃 𝒄 𝒃𝟐 𝒄𝟐 𝒃𝒄[(𝒃 − 𝒄)𝟐 + 𝟐𝒃𝒄 − 𝒂𝟐 ]
= + −
(𝒂 + 𝒄)𝟐 (𝒂 + 𝒃)𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄)
𝟐
𝟏 𝟏 𝒃𝒄(𝒃 − 𝒄)𝟐 𝒂𝟐 𝒃𝒄
= 𝒃𝟐 𝒄𝟐 ( − ) − + =
𝒂+𝒄 𝒂+𝒃 (𝒂 + 𝒃)(𝒂 + 𝒄) (𝒂 + 𝒃)(𝒂 + 𝒄)
𝒃𝟐 𝒄𝟐 (𝒃 − 𝒄)𝟐 𝒃𝒄(𝒃 − 𝒄)𝟐 𝒂𝟐 𝒃𝒄
= − +
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄) (𝒂 + 𝒃)(𝒂 + 𝒄)
𝒃𝒄(𝒃 − 𝒄)𝟐 [(𝒂 + 𝒃)(𝒂 + 𝒄) − 𝒃𝒄] 𝒂𝟐 𝒃𝒄
=− + =
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄)
𝟐𝒔𝒂𝒃𝒄(𝒃 − 𝒄)𝟐 𝒂𝟐 𝒃𝒄 𝒂𝟐 𝒃𝒄
=− + ≤ → 𝑩𝟏 𝑪𝟏
(𝒂 + 𝒃)𝟐 (𝒂 + 𝒄)𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄) (𝒂 + 𝒃)(𝒂 + 𝒄)
𝒃𝒄
≤ 𝒂√ ( 𝟏)
(𝒂 + 𝒃)(𝒂 + 𝒄)
𝟐[𝑨𝑩𝟏 𝑪𝟏 ] 𝟏 𝒃𝒄. 𝑭
𝑵𝒐𝒘, 𝒘𝒆 𝒉𝒂𝒗𝒆 ∶ 𝒉𝑨 = 𝒘𝒊𝒕𝒉 [𝑨𝑩𝟏 𝑪𝟏 ] = . 𝑨𝑩𝟏 . 𝑨𝑪𝟏 . 𝐬𝐢𝐧 𝑨 =
𝑩𝟏 𝑪𝟏 𝟐 (𝒂 + 𝒃)(𝒂 + 𝒄)
(𝟏)
𝟐𝒃𝒄. 𝑭 𝟐𝒃𝒄. 𝑭 √(𝒂 + 𝒃)(𝒂 + 𝒄)
→ 𝒉𝑨 = ≥⏞ .
(𝒂 + 𝒃)(𝒂 + 𝒄). 𝑩𝟏 𝑪𝟏 (𝒂 + 𝒃)(𝒂 + 𝒄) 𝒂√𝒃𝒄
𝒃𝒄
= 𝒉𝒂 . √
(𝒂 + 𝒃)(𝒂 + 𝒄)

www.ssmrmh.ro
𝒉𝒂 (𝒂 + 𝒃)(𝒂 + 𝒄) 𝟑 𝒉𝒂 𝟔 (𝒂 + 𝒃)(𝒂 + 𝒄)
→ ≤√ → √ ≤√
𝒉𝑨 𝒃𝒄 𝒉𝑨 𝒃𝒄
𝑨𝑴−𝑮𝑴 𝟑
𝟑
𝟔 𝒂+𝒃 𝒂+𝒄 √𝟐 𝒂 + 𝒃 𝒂 + 𝒄
= √𝟐. √ . . 𝟏. 𝟏. 𝟏. 𝟏 ⏞
≤ ( + + 𝟒. 𝟏)
𝟐𝒄 𝟐𝒃 𝟔 𝟐𝒄 𝟐𝒃
𝟑
𝒉𝒂 √𝟐 𝟑 𝒂+𝒃 𝒂+𝒄
→∑√ ≤ ∑( + + 𝟒)
𝒉𝑨 𝟔 𝟐𝒄 𝟐𝒃
𝑬𝒖𝒍𝒆𝒓
𝟑 𝑩𝒂𝒏𝒅𝒊𝒍𝒂 𝟑
√𝟐 𝒂 𝒃 √𝟐 𝑹 𝑹 𝟑 𝑹
= (𝟏𝟐 + ∑ ( + )) ⏞
≤ (𝟏𝟐. + ∑ ) = 𝟑√𝟐.
𝟔 𝒃 𝒂 𝟔 𝟐𝒓 𝒓 𝟐𝒓
𝟑 𝒉𝒂 𝟑 𝑹
𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆, ∑√ ≤ 𝟑√𝟐.
𝒉𝑨 𝟐𝒓
𝒄𝒚𝒄
JP.397 Solve for real numbers:
𝒙 −𝟏) 𝒙 +𝟏)
𝟑𝐥𝐨𝐠𝟐(𝟑 = 𝟐𝐥𝐨𝐠𝟑(𝟐 +𝟏
Proposed by Ionuț Florin Voinea-Romania
𝒇(𝒙) = 𝒈(𝒚)
Lemma. Let be 𝒇, 𝒈: (𝟎, ∞) → ℝ have strictly monotonically and {𝒇(𝒚) = 𝒈(𝒛).
𝒇(𝒛) = 𝒈(𝒙)
Then 𝒙 = 𝒚 = 𝒛 has unique solution for the system.
Proof. Let’s suppose 𝒇, 𝒈 − increasing functions.
If 𝒙 < 𝑦 then 𝒇(𝒙) < 𝑓 (𝒚) ⇒ 𝒈(𝒚) < 𝑔(𝒛), 𝒈 ↗⇒ 𝒚 < 𝑧, 𝑓 ↗⇒ 𝑓(𝒚) < 𝑓(𝒛) ⇒
𝒈(𝒛) < 𝑔(𝒙), 𝒈 ↗⇒ 𝒛 < 𝑥.
So, 𝒙 < 𝑦 < 𝑧 < 𝑥 impossible. Analogous, for 𝒙 > 𝑦 ⇒ 𝑥 > 𝑦 > 𝑧 > 𝑥 impossible.
Hence, 𝒙 = 𝒚 ⇒ 𝒈(𝒚) = 𝒈(𝒛); 𝒈 ↗⇒ 𝒈 −injective function, then 𝒚 = 𝒛 ⇒ 𝒙 = 𝒚 = 𝒛.
𝟑𝒙 − 𝟏 > 0 ⇒ 𝟑𝒙 > 1 ⇒ 𝑥 ∈ (𝟎, ∞)
𝒙 −𝟏) 𝒙 +𝟏) 𝒙 +𝟏)
𝟑𝐥𝐨𝐠𝟐 (𝟑 = 𝟐𝐥𝐨𝐠 𝟑 (𝟐 + 𝟏 ⟺ 𝐥𝐨𝐠 𝟐 (𝟑𝒙 − 𝟏) = 𝐥𝐨𝐠 𝟑 ( 𝟐𝐥𝐨𝐠 𝟑 (𝟐 + 𝟏); (∗)
𝒙 +𝟏)
Let 𝒚 = 𝐥𝐨𝐠 𝟐 (𝟑𝒙 − 𝟏) and 𝒛 = 𝐥𝐨𝐠 𝟑 ( 𝟐𝐥𝐨𝐠 𝟑 (𝟐 + 𝟏) = 𝐥𝐨𝐠 𝟑 (𝟐𝒙 + 𝟏) then (∗) ⟺
𝐥𝐨𝐠 𝟐 (𝟐𝒙 + 𝟏) = 𝐥𝐨𝐠 𝟑 (𝟐𝒚 + 𝟏) = 𝒛 ⟺ 𝐥𝐨𝐠 𝟐 (𝟑𝒙 − 𝟏) = 𝒛 ⟺ 𝟑𝒙 − 𝟏 = 𝟐𝒛
𝟑𝒙 = 𝟐𝒛 + 𝟏 ⟺ 𝒙 = 𝐥𝐨𝐠 𝟑 (𝟐𝒛 + 𝟏)

www.ssmrmh.ro
𝒙 = 𝐥𝐨𝐠 𝟑 (𝟐𝒛 + 𝟏) 𝟑𝒙 = 𝟐𝒛 + 𝟏
𝒙
So, {𝒚 = 𝐥𝐨𝐠 𝟑 (𝟐 + 𝟏) ⟺ {𝟑𝒚 = 𝟐𝒙 + 𝟏
𝒛 = 𝐥𝐨𝐠 𝟑 (𝟐𝒚 + 𝟏) 𝟑𝒛 = 𝟐𝒚 + 𝟏
Let 𝒇: (𝟎, ∞) → ℝ, 𝒇(𝒙) = 𝟑𝒙 , 𝒇 ↗; 𝒈: (𝟎, ∞) → ℝ, 𝒈(𝒛) = 𝟐𝒛 + 𝟏, 𝒈 ↗, from Lemma it
follows that 𝒙 = 𝒚 = 𝒛.
𝟐 𝒙 𝟏 𝒙
𝒙 = 𝒛 ⇒ 𝟑𝒙 = 𝟐𝒙 + 𝟏 ⟺ (𝟑) + (𝟑) = 𝟏 ⇒ 𝒙 = 𝟏 solution.
Solution 2 by Florentin Vişescu-Romania

𝟑𝒙 − 𝟏 > 0 ⇒ 𝑥 > 0.Let 𝒇: (𝟎, ∞) → ℝ, 𝒇(𝒙) = 𝐥𝐨𝐠 𝟐 (𝟑𝒙 − 𝟏) increasing, then 𝒇 −injective.
Let 𝒚 ∈ ℝ ⇒ ∃𝒙 > 0 such that 𝒇(𝒙) = 𝒚 ⇔ 𝐥𝐨𝐠 𝟐 (𝟑𝒙 − 𝟏) = 𝒚 ⇔ 𝟑𝒙 = 𝟐𝒚 + 𝟏 ⇔
𝒙 = 𝐥𝐨𝐠 𝟑 (𝟐𝒚 + 𝟏). So, 𝒇 −surjective. Hence, 𝒇 −invertible with 𝒇−𝟏 (𝒙) = 𝐥𝐨𝐠 𝟑 (𝟐𝒙 + 𝟏)
So, we have:
−𝟏(𝒙) −𝟏 (𝒙) −𝟏(𝒇−𝟏 (𝒙))
𝟑𝒇(𝒙) = 𝟐𝒇 + 𝟏 ⇒ 𝒇(𝒙) = 𝐥𝐨𝐠 𝟑 (𝟐𝒇 + 𝟏) ⇒ 𝒙 = 𝐥𝐨𝐠 𝟑 (𝟐𝒇 + 𝟏)
−𝟏 (𝒇−𝟏 (𝒙)) −𝟏(𝒇−𝟏 (𝒙))
⇒ 𝟑𝒙 = 𝟐𝒇 + 𝟏 ⇒ 𝟑𝒙 − 𝟏 = 𝟐𝒇
⇒ 𝐥𝐨𝐠 𝟐 (𝟑𝒙 − 𝟏) = 𝒇−𝟏 (𝒇−𝟏 (𝒙)) ⇒ 𝒇 (𝒇(𝒇(𝒙))) = 𝒙
Lemma. If 𝒇 is strictly increasing function with 𝒇 (𝒇(𝒇(𝒙))) = 𝒂, then 𝒇(𝒂) = 𝒂.
Proof. Let’s suppose 𝒇(𝒂) = 𝒃 > 𝑎. Then 𝒇(𝒇(𝒂)) = 𝒇(𝒃) > 𝑓(𝒂) = 𝒂 and
𝒇 (𝒇(𝒇(𝒂))) > 𝑓 (𝒂) > 𝑎 impossible!
Let’s suppose 𝒇(𝒂) = 𝒃 < 𝑎. Then 𝒇(𝒇(𝒂)) = 𝒇(𝒃) < 𝑓(𝒂) < 𝑎 and
𝒇 (𝒇(𝒇(𝒂))) < 𝑓 (𝒂) < 𝑎 impossible!
Return to the problem. 𝒇 (𝒇(𝒇(𝒙))) = 𝒙 ⇒ 𝒇(𝒙) = 𝒙 ⇒ 𝐥𝐨𝐠 𝟐 (𝟑𝒙 − 𝟏) = 𝒙 ⇒ 𝟑𝒙 = 𝟐𝒙 + 𝟏
𝟑 𝒙 𝟏 𝒙 𝟑 𝒙 𝟏 𝒙
( ) − ( ) = 𝟏; (𝒙 → ( ) ↗; 𝒙 → ( ) ↘)
𝟐 𝟐 𝟐 𝟐
𝟑 𝒙 𝟏 𝒙
So, 𝒖(𝒙) = (𝟐) − (𝟐) injective and 𝒙 = 𝟏 unique solution.
Solution 3 by Agayev Seddredin-Azerbaijan

𝒙 −𝟏) 𝒙 +𝟏)
𝟑𝐥𝐨𝐠𝟐 (𝟑 = 𝟐𝐥𝐨𝐠 𝟑 (𝟐 +𝟏

www.ssmrmh.ro
𝐥𝐨𝐠 𝟐 (𝟑𝒙 − 𝟏) = 𝒂 𝟑𝒙 = 𝟐𝒂 + 𝟏
{ ⇒ {
𝐥𝐨𝐠 𝟑 (𝟐𝒙 + 𝟏) = 𝒃 𝟑𝒃 = 𝟐𝒙 + 𝟏
We have: 𝟑𝒂 = 𝟐𝒃 + 𝟏
If 𝒂 ≥ 𝒃 ≥ 𝒙; (∗) ⇒ 𝟐𝒂 + 𝟏 = 𝟑𝒙 ≤ 𝟑𝒃 = 𝟐𝒙 + 𝟏 ⇒ 𝒂 ≤ 𝒙; (∗∗)
From (∗), (∗∗) we get 𝒙 = 𝒂.
𝟐 𝒙 𝟏 𝒙
𝟑𝒙 = 𝟐𝒙 + 𝟏 ⇒ ( ) + ( ) = 𝟏
𝟑 𝟑
𝟐 𝒙 𝟐 𝟏 𝟏 𝒙 𝟏 𝟏 𝟐 𝒙 𝟏 𝒙
𝒙 > 1 ⇒ ( ) < ( ) , ( ) < ( ) ⇒ ( ) + ( ) < 1; (𝒏𝒐 𝒔𝒐𝒍𝒖𝒕𝒊𝒐𝒏!)
𝟑 𝟑 𝟑 𝟑 𝟑 𝟑
𝟐 𝒙 𝟏 𝒙
𝒙 < 1 ⇒ ( ) + ( ) > 1; (𝒏𝒐 𝒔𝒐𝒍𝒖𝒕𝒊𝒐𝒏!)
𝟑 𝟑
Therefore, 𝒙 = 𝟏 is unique solution.
Solution 4 by Jamal Issah-Ghana
𝒙 −𝟏) 𝒙 +𝟏)
𝟑𝐥𝐨𝐠𝟐 (𝟑 = 𝟐𝐥𝐨𝐠 𝟑 (𝟐 +𝟏
𝐥𝐨𝐠 (𝟑𝒙 − 𝟏) = 𝒎 𝟐𝒎 = 𝟑𝒙 − 𝟏
{ 𝟐 𝒙 ⇒{ 𝒏
𝐥𝐨𝐠 𝟑 (𝟐 + 𝟏) = 𝒏 𝟑 = 𝟐𝒙 + 𝟏
𝒎 𝒏 𝒎 𝒏
We have: 𝟑𝟐 = 𝟐𝟑 + 𝟏 ⇔ 𝟑𝟐 − 𝟐𝟑 = 𝟏
From Catalan’s conjecture: 𝒎 = 𝒏 = 𝟏. Hence, 𝟑𝒙 − 𝟏 = 𝟐 ⇔ 𝒙 = 𝟏.
Solution 5 by Hikmat Mammmadov-Azerbaijan
∵ 𝒂𝐥𝐨𝐠𝒃 𝒄 = 𝒄𝐥𝐨𝐠𝐛 𝒂 ⇒ (𝟑𝒙 − 𝟏)𝐥𝐨𝐠𝟐 𝟑 = (𝟐𝒙 + 𝟏)𝐥𝐨𝐠𝟑 𝟐 + 𝟏
𝒙
{𝟑𝒙 − 𝟏 = 𝒗 ⇒ 𝟐𝒗⋅𝐥𝐨𝐠𝟐 𝟑 + 𝟑𝒖⋅𝐥𝐨𝐠𝟑 𝟐 + 𝟏 = 𝟎
𝟐 +𝟏 =𝒖
𝟑𝒗 = 𝟐𝒖 + 𝟏 ⇔ 𝟑𝒗 − 𝟐𝒖 = 𝟏 ⇒ 𝒖 = 𝒗 = 𝟏 ⇒ 𝒙 = 𝟏
Solution 6 by Khaled Abd Imouti-Damascus-Syria
𝒏 + 𝟏 = 𝒏 + 𝟏 ⇒ 𝒆𝐥𝐨𝐠(𝒏+𝟏) = 𝒆𝐥𝐨𝐠 𝒏 + 𝟏
𝐥𝐨𝐠(𝒏+𝟏) 𝐥𝐨𝐠 𝒏
⋅𝐥𝐨𝐠(𝒏+𝟏−𝟏) ⋅𝐥𝐨𝐠(𝒏+𝟏)
𝒆 𝐥𝐨𝐠 𝒏 = 𝒆 𝐥𝐨𝐠(𝒏+𝟏)
𝐥𝐨𝐠(𝒏+𝟏) 𝐥𝐨𝐠 𝒏
⋅𝐥𝐨𝐠(𝒆𝐥𝐨𝐠(𝒏+𝟏)−𝟏) ⋅𝐥𝐨𝐠(𝒆𝐥𝐨𝐠 𝒏 +𝟏)
𝒆 𝐥𝐨𝐠 𝒏 = 𝒆𝐥𝐨𝐠(𝒏+𝟏) +𝟏
𝐥𝐨𝐠(𝒏+𝟏)
𝒙 −𝟏) 𝐥𝐨𝐠(𝒆𝒙⋅𝐥𝐨𝐠(𝒏+𝟏)−𝟏 )⋅
(𝒏 + 𝟏)𝐥𝐨𝐠𝒏 ((𝒏+𝟏) =𝒆 𝐥𝐨𝐠 𝒏

www.ssmrmh.ro
𝐥𝐨𝐠 𝒏
𝐥𝐨𝐠 𝐧+𝟏 (𝒏𝒙 +𝟏) 𝐥𝐨𝐠(𝒆𝒙 𝐥𝐨𝐠 𝒏 +𝟏)⋅
𝒏 =𝒆 𝐥𝐨𝐠(𝒏+𝟏) =𝟏⇒𝒙=𝟏
JP.398 If 𝒂, 𝒃, 𝒄 > 0; 𝑎𝑏√𝒂𝒃 + 𝒃𝒄√𝒃𝒄 + 𝒄𝒂√𝒄𝒂 = 𝟑 then:

𝟐
∑(𝒂√𝒂 − 𝒃√𝒃) + 𝟐𝟒 ≤ ∑(𝒂 + 𝒃)𝟑

First, we prove that:
𝟐
(𝒂√𝒂 − 𝒃√𝒃) + 𝟖𝒂𝒃√𝒂𝒃 ≤ (𝒂 + 𝒃)𝟑 ; (𝟏)
𝒂𝟑 + 𝒃𝟑 − 𝟐𝒂𝒃√𝒂𝒃 + 𝟖𝒂𝒃√𝒂𝒃 ≤ 𝒂𝟑 + 𝒃𝟑 + 𝟑𝒂𝒃(𝒂 + 𝒃) ⇔

𝟑𝒂𝒃(𝒂 + 𝒃) − 𝟔𝒂𝒃√𝒂𝒃 ≥ 𝟎 ⇔ 𝟑𝒂𝒃(𝒂 + 𝒃 − 𝟐√𝒂𝒃) ≥ 𝟎
𝟐
⇔ 𝟑𝒂𝒃(√𝒂 − √𝒃) ≥ 𝟎
By adding in (1), we get:
𝟐
∑(𝒂√𝒂 − 𝒃√𝒃) + 𝟖 ∑ 𝒂𝒃√𝒂𝒃 ≤ ∑(𝒂 + 𝒃)𝟑 ⇔
𝒄𝒚𝒄 𝒄𝒚𝒄 𝒄𝒚𝒄
𝟐 𝟐
∑(𝒂√𝒂 − 𝒃√𝒃) + 𝟖 ⋅ 𝟑 ≤ ∑(𝒂 + 𝒃)𝟑 ⇔ ∑(𝒂√𝒂 − 𝒃√𝒃) + 𝟐𝟒 ≤ ∑(𝒂 + 𝒃)𝟑
𝒄𝒚𝒄 𝒄𝒚𝒄 𝒄𝒚𝒄 𝒄𝒚𝒄
Equality holds for 𝒂 = 𝒃 = 𝒄.

Solution 2 by George Florin Şerban-Romania
𝟐
∑(𝒂√𝒂 − 𝒃√𝒃) + 𝟐𝟒 = ∑(𝒂𝟑 + 𝒃𝟑 − 𝟐𝒂𝒃√𝒂𝒃) + 𝟐𝟒 =
(𝟏)
= ∑(𝒂𝟑 + 𝒃𝟑 ) − 𝟐 ∑ 𝒂𝒃√𝒂𝒃 + 𝟐𝟒 = 𝟐 ∑ 𝒂𝟑 + 𝟏𝟖 ≤ ∑(𝒂 + 𝒃)𝟑 =
= 𝟐 ∑ 𝒂𝟑 + 𝟑 ∑ 𝒂𝒃(𝒂 + 𝒃)
(𝟏) ⇔ ∑ 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟔;
𝒄𝒚𝒄

www.ssmrmh.ro
𝑨𝑴−𝑮𝑴
∑ 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟐 ∑ 𝒂𝒃√𝒂𝒃 = 𝟔
Therefore,
𝟐
∑(𝒂√𝒂 − 𝒃√𝒃) + 𝟐𝟒 ≤ ∑(𝒂 + 𝒃)𝟑

Solution 3 by Fayssal Abdelli-Bejaia-Algerie
∑(𝒂 + 𝒃)𝟑 = ∑(𝒂𝟑 + 𝒃𝟑 + 𝟑𝒂𝟐 𝒃 + 𝟑𝒂𝒃𝟐 ) = 𝟐 ∑ 𝒂𝟑 + 𝟑 ∑ 𝒂𝒃(𝒂 + 𝒃) ; (𝟏)

𝟐
∑(𝒂√𝒂 − 𝒃√𝒃) = ∑(𝒂𝟑 + 𝒃𝟑 − 𝟐𝒂𝒃√𝒂𝒃) = 𝟐 ∑ 𝒂𝟑 − 𝟐 ∑ 𝒂𝒃√𝒂𝒃 =
= 𝟐(𝒂𝟑 + 𝒃𝟑 + 𝒄𝟑 ) − 𝟔
Hence,
𝟐
𝟐𝟒 + ∑(𝒂√𝒂 − 𝒃√𝒃) = 𝟏𝟖 + 𝟐(𝒂𝟑 + 𝒃𝟑 + 𝒄𝟑 ); (𝟐)
𝒄𝒚𝒄
We must to prove:
𝟑 ∑ 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟏𝟖 ⇔ ∑ 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟔; (𝟑)

𝑨𝑴−𝑮𝑴
∑ 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟐 ∑ 𝒂𝒃√𝒂𝒃 ⇒ (𝟑)𝐭𝐫𝐮𝐞 ⇒ (𝟐) 𝐭𝐫𝐮𝐞.
Therefore,
𝟐
∑(𝒂√𝒂 − 𝒃√𝒃) + 𝟐𝟒 ≤ ∑(𝒂 + 𝒃)𝟑

We have:
𝟐
∑(𝒂√𝒂 − 𝒃√𝒃) + 𝟐𝟒 = 𝟐(𝒂𝟑 + 𝒃𝟑 + 𝒄𝟑 ) + 𝟏𝟖
𝒄𝒚𝒄

www.ssmrmh.ro
𝟐(𝒂𝟑 + 𝒃𝟑 + 𝒄𝟑 ) + 𝟏𝟖 ≤ ∑(𝒂 + 𝒃)𝟑 ⇔
𝒄𝒚𝒄
𝑨𝑴−𝑮𝑴
𝟑 ∑ 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟏𝟖 ⇔ ∑ 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟔, ∑ 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟐 ∑ 𝒂𝒃√𝒂𝒃
Therefore,
𝟐
∑(𝒂√𝒂 − 𝒃√𝒃) + 𝟐𝟒 ≤ ∑(𝒂 + 𝒃)𝟑

𝟏
𝒂𝟒 + 𝒃𝟒 + 𝒄𝟒 ≥ 𝟏𝟔𝑭𝟐 + ((𝒂𝟐 − 𝒃𝟐 )𝟐 + (𝒃𝟐 − 𝒄𝟐 )𝟐 + (𝒄𝟐 − 𝒂𝟐 )𝟐 )
𝟐
Proposed by D.M. Bătinețu-Giurgiu, Daniel Sitaru-Romania
Solution 1 by proposers
∑(𝒂𝟐 − 𝒃𝟐 )𝟐 = 𝟐(𝒂𝟒 + 𝒃𝟒 + 𝒄𝟒 ) − 𝟐 ∑ 𝒂𝟐 𝒃𝟐
𝟏
𝒂𝟒 + 𝒃𝟒 + 𝒄𝟒 = ∑ 𝒂𝟐 𝒃𝟐 + ∑(𝒂𝟐 − 𝒃𝟐 )𝟐 =
𝟐
(𝒂𝒃)𝟐 (𝒃𝒄)𝟐 (𝒄𝒂)𝟐 𝟏

= + + + ∑(𝒂𝟐 − 𝒃𝟐 )𝟐 ≥
𝟏 𝟏 𝟏 𝟐
𝒄𝒚𝒄
𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎 (𝒂𝒃 + 𝒃𝒄 + 𝒄𝒂)𝟐 𝟏

≥ + ∑(𝒂𝟐 − 𝒃𝟐 )𝟐 ≥
𝟏+𝟏+𝟏 𝟐
𝒄𝒚𝒄
𝟐
𝑮𝑶𝑹𝑫𝑶𝑵 (𝟒√𝟑𝑭) 𝟏 𝟏
≥ + ∑(𝒂𝟐 − 𝒃𝟐 )𝟐 = 𝟏𝟔𝑭𝟐 + ∑(𝒂𝟐 − 𝒃𝟐 )𝟐
𝟑 𝟐 𝟐

𝟐(𝒂𝟒 + 𝒃𝟒 + 𝒄𝟒 ) − (𝒂𝟐 − 𝒃𝟐 )𝟐 − (𝒃𝟐 − 𝒄𝟐 )𝟐 − (𝒄𝟐 − 𝒂𝟐 )𝟐 =
= 𝒂𝟒 + 𝒃𝟒 − (𝒂𝟐 − 𝒃𝟐 )𝟐 + 𝒃𝟒 + 𝒄𝟒 − (𝒃𝟐 − 𝒄𝟐 )𝟐 + 𝒄𝟒 + 𝒂𝟒 − (𝒄𝟐 − 𝒂𝟐 )𝟐 =
= 𝒂𝟒 + 𝒂𝟐 (𝟐𝒄𝟐 − 𝒂𝟐 ) + 𝒃𝟒 + 𝒃𝟐 (𝟐𝒂𝟐 − 𝒃𝟐 ) + 𝒄𝟒 + 𝒄𝟐 (𝟐𝒃𝟐 − 𝒄𝟐 ) =
𝟐 𝟐 𝟐
= 𝟐(𝒂𝒃)𝟐 + 𝟐(𝒃𝒄)𝟐 + 𝟐(𝒄𝒂)𝟐 ≥ (𝒂𝒃 + 𝒃𝒄 + 𝒄𝒂)𝟐 ≥ (𝟒√𝟑𝑭) = 𝟑𝟐𝑭𝟐
𝟑 𝟑
www.ssmrmh.ro
⇒ 𝟐 𝒂 + 𝒃 + 𝒄 ≥ 𝟑𝟐𝑭𝟐 + (𝒂𝟐 − 𝒃𝟐 )𝟐 + (𝒃𝟐 − 𝒄𝟐 )𝟐 + (𝒄𝟐 − 𝒂𝟐 )𝟐
( 𝟒 𝟒 𝟒)
𝟏
⇒ 𝒂𝟒 + 𝒃𝟒 + 𝒄𝟒 ≥ 𝟏𝟔𝑭𝟐 + [(𝒂𝟐 − 𝒃𝟐 )𝟐 + (𝒃𝟐 − 𝒄𝟐 )𝟐 + (𝒄𝟐 − 𝒂𝟐 )𝟐 ]
𝟐
Solution 3 by Marian Ursărescu-Romania
We must show that:
𝟏
𝒂𝟒 + 𝒃𝟒 + 𝒄𝟒 ≥ 𝟏𝟔𝑭𝟐 + (𝟐𝒂𝟒 + 𝟐𝒃𝟒 + 𝟐𝒄𝟒 − 𝟐𝒂𝟐 𝒃𝟐 − 𝟐𝒃𝟐 𝒄𝟐 − 𝟐𝒄𝟐 𝒂𝟐 ) ⇔
𝟐
𝒂𝟐 𝒃𝟐 + 𝒃𝟐 𝒄𝟐 + 𝒄𝟐 𝒂𝟐 ≥ 𝟏𝟔𝑭𝟐 ; (𝟏)
But (𝒂𝒃)𝟐 + (𝒃𝒄)𝟐 + (𝒄𝒂)𝟐 ≥ 𝒂𝒃𝒄(𝒂 + 𝒃 + 𝒄); (𝟐)
From (1)&(2) we must show that: 𝒂𝒃𝒄(𝒂 + 𝒃 + 𝒄) ≥ 𝟏𝟔𝑭𝟐 ; (𝟑)
But 𝒂𝒃𝒄 = 𝟒𝑹𝒓𝒔, 𝒂 + 𝒃 + 𝒄 = 𝟐𝒔, 𝑭 = 𝒔𝒓; (𝟒)
From (3)&(4) we must show that: 𝟒𝑹𝒓𝒔 ⋅ 𝟐𝒔 ≥ 𝟏𝟔𝒓𝟐 𝒔𝟐 ⇔ 𝑹 ≥ 𝟐𝒓 (𝑬𝒖𝒍𝒆𝒓).
Solution 4 by Avishek Mitra-West Bengal-India
𝟏 𝟏
∑ 𝒂𝟒 = 𝟏𝟔𝑭𝟐 + ∑(𝒂𝟐 − 𝒃𝟐 )𝟐 ⇔ ∑ 𝒂𝟒 ≥ 𝟏𝟔𝑭𝟐 + (𝟐 ∑ 𝒂𝟒 − 𝟐 ∑ 𝒂𝟐 𝒃𝟐 )
𝟐 𝟐
𝒄𝒚𝒄 𝒄𝒚𝒄 𝒄𝒚𝒄 𝒄𝒚𝒄 𝒄𝒚𝒄
∑ 𝒂𝟐 𝒃𝟐 ≥ 𝟏𝟔𝑭𝟐 (𝒏𝒆𝒆𝒅 𝒕𝒐 𝒑𝒓𝒐𝒗𝒆)

𝒄𝒚𝒄
𝟐
𝑪𝑩𝑺
∵ (𝟏𝟐 + 𝟏𝟐 + 𝟏𝟐 ) (∑ 𝒂𝟐 𝒃𝟐 ) ≥ (∑ 𝒂𝒃) ⇒
𝟐
(𝒔𝟐 + 𝒓𝟐 + 𝟒𝑹𝒓)𝟐
𝟐 𝟐
𝑮𝒆𝒓𝒓𝒆𝒕𝒔𝒆𝒏 𝟏𝟔𝑹𝒓 − 𝟓𝒓𝟐 + 𝒓𝟐 + 𝟒𝑹𝒓
∑𝒂 𝒃 ≥ ≥ ( ) =
𝟑 𝟑
𝒄𝒚𝒄
𝟐𝒔 𝟐 𝒔 𝟐 𝟑𝟔𝑭 𝟐
𝟐 𝟐𝟎 ⋅ 𝒓 + 𝒓 − 𝟒𝒓 ⋅ ( )
𝟐𝟎𝑹𝒓 − 𝟒𝒓𝟐 𝑴𝒊𝒕𝒓𝒊𝒏𝒐𝒗𝒊𝒄 𝟑√𝟑 𝟑√𝟑 𝟑√𝟑
=( ) ≥ ( ) = = 𝟏𝟔𝑭𝟐
𝟑 𝟑 𝟑
𝟑
JP.400 If 𝒙, 𝒚, 𝒛 > 0; 𝟑√𝒙𝒚 + 𝟑√𝒚𝒛 + √𝒛𝒙 = 𝟑 then:
𝟑 𝟑 𝟑 𝟑
√𝒙(𝒚 + 𝒛) + √𝒚(𝒛 + 𝒙) + √𝒛(𝒙 + 𝒚) ≥ 𝟑 √𝟐
First we prove that if 𝒙, 𝒚 > 0 then 𝟑√𝟒(𝒙 + 𝒚) ≥ 𝟑√𝒙 + 𝟑√𝒚; (𝟏)
𝟑
Denote 𝒙 = 𝒂𝟑 , 𝒚 = 𝒃𝟑 we have: √𝟒(𝒂𝟑 + 𝒃𝟑 ) ≥ 𝒂 + 𝒃 ⇔

www.ssmrmh.ro
𝟒 𝒂 + 𝒃 ≥ 𝒂 + 𝒃 ⇔ 𝟒𝒂𝟑 + 𝟒𝒃𝟑 ≥ 𝒂𝟑 + 𝒃𝟑 + 𝟑𝒂𝒃(𝒂 + 𝒃) ⇔
( 𝟑 𝟑) ( ) 𝟑
𝟑(𝒂𝟑 − 𝒃𝟑 ) − 𝟑𝒂𝒃(𝒂 + 𝒃) ≥ 𝟎 ⇔ (𝒂 + 𝒃)(𝒂𝟐 − 𝒂𝒃 + 𝒃𝟐 ) − 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟎

⇔ (𝒂 + 𝒃)(𝒂𝟐 − 𝟐𝒂𝒃 + 𝒃𝟐 ) ≥ 𝟎 ⇔ (𝒂 + 𝒃)(𝒂 − 𝒃)𝟐 ≥ 𝟎, which is true for all 𝒂, 𝒃 ≥ 𝟎.
Equality holds for 𝒂 = 𝒃 ⇔ 𝒙 = 𝒚.
Multiplying (1) with 𝟑√𝒛, it follows that:
𝟑
√𝟒𝒛(𝒙 + 𝒚) ≥ 𝟑√𝒙𝒛 + 𝟑√𝒚𝒛; (𝟐)
Analogous,
𝟑
√𝟒𝒙(𝒚 + 𝒛) ≥ 𝟑√𝒙𝒚 + 𝟑√𝒙𝒛; (𝟑)
𝟑
√𝟒𝒚(𝒛 + 𝒙) ≥ 𝟑√𝒛𝒚 + 𝟑√𝒙𝒚; (𝟒)
By adding (2),(3),(4) we obtain:
𝟑 𝟑 𝟑 𝟑 𝟑
√𝟒 (√𝒙(𝒚 + 𝒛) + √𝒚(𝒛 + 𝒙) + √𝒛(𝒙 + 𝒚)) ≥ 𝟐( 𝟑√𝒙𝒚 + 𝟑√𝒚𝒛 + √𝒛𝒙) = 𝟐 ⋅ 𝟑 = 𝟔
Hence,
𝟑
𝟑 𝟑 𝟑 𝟔 𝟔√𝟐 𝟑
√𝒙(𝒚 + 𝒛) + √𝒚(𝒛 + 𝒙) + √𝒛(𝒙 + 𝒚) ≥ 𝟑 = 𝟑 = 𝟑√𝟐
√𝟒 √𝟖
Equality holds for 𝒙 = 𝒚 = 𝒛 = 𝟏.
𝟑 𝟑 𝟑 𝟑
√𝒙(𝒚 + 𝒛) + √𝒚(𝒛 + 𝒙) + √𝒛(𝒙 + 𝒚) ≥ 𝟑 √ 𝟑√𝒙𝒚𝒛(𝒙 + 𝒚)(𝒚 + 𝒛)(𝒛 + 𝒙)
𝑨𝑴−𝑮𝑴
But: (𝒙 + 𝒚)(𝒚 + 𝒛)(𝒛 + 𝒙) ≥ 𝟖𝒙𝒚𝒛 (𝑪𝒆𝒔𝒂𝒓𝒐); (𝟏)
𝟑 𝟑 𝟑 𝟑 𝟑 𝟑 𝟑
√𝒙(𝒚 + 𝒛) + √𝒚(𝒛 + 𝒙) + √𝒛(𝒙 + 𝒚) ≥ 𝟑 √ √𝟖(𝒙𝒚𝒛)𝟐 = 𝟑 √𝟐√(𝒙𝒚𝒛)𝟐
𝟑 𝟑 𝟔
But 𝟑√𝒙𝒚 + 𝟑√𝒚𝒛 + 𝟑√𝒛𝒙 ≥= 𝟑 √ √(𝒙𝒚𝒛)𝟐 ⇒ 𝟑 ≥ 𝟑√(𝒙𝒚𝒛)𝟐 ⇒ 𝒙𝒚𝒛 ≤ 𝟏; (𝟐)
𝟑 𝟑 𝟑 𝟑 𝟑
? 𝟑
√𝒙(𝒚 + 𝒛) + √𝒚(𝒛 + 𝒙) + √𝒛(𝒙 + 𝒚) ≥ 𝟑 √𝟐 √(𝒙𝒚𝒛)𝟐 ≥ 𝟑√𝟐 ⇒
𝟑 𝟔 𝟑
√𝟐 ⋅ √(𝒙𝒚𝒛)𝟐 ≥ √𝟐 ⇒ 𝒙𝒚𝒛 ≥ 𝟏; (𝟑)
𝟑 𝟑 𝟑 𝟑 𝟑 𝟑
⇒ √𝒙(𝒚 + 𝒛) + √𝒚(𝒛 + 𝒙) + √𝒛(𝒙 + 𝒚) ≥ 𝟑 √𝟐 √(𝒙𝒚𝒛)𝟐 ≥ 𝟑√𝟐

www.ssmrmh.ro
𝟑 𝟑 𝟑
√𝒙(𝒚 + 𝒛) + √𝒚(𝒛 + 𝒙) + √𝒛(𝒙 + 𝒚) =
𝟑 𝟏 𝟑 𝟏 𝟑 𝟏
= √ (𝟏 + 𝟏)(𝒙 + 𝒙)(𝒚 + 𝒛) + √ (𝟏 + 𝟏)(𝒚 + 𝒚)(𝒛 + 𝒙) + √ (𝟏 + 𝟏)(𝒛 + 𝒛)(𝒙 + 𝒚)
𝟒 𝟒 𝟒
𝟑 𝟏
𝟑 𝟑
≥ √ ( 𝟑√𝟏 ⋅ 𝒙𝒚 + 𝟑√𝟏 ⋅ 𝒚𝒛 + 𝟑√𝟏 ⋅ 𝒚𝒙 + √𝟏 ⋅ 𝒛𝒙 + 𝟑√𝟏 ⋅ 𝒛𝒚 + √𝟏 ⋅ 𝒙𝒛) =
𝟒
𝟑 𝟏 𝟑 𝟏
𝟑
≥ 𝟐 ⋅ √ ( 𝟑√𝒙𝒚 + 𝟑√𝒚𝒛 + 𝟑√𝒛𝒙) = 𝟐 ⋅ √ ⋅ 𝟑 = 𝟑√𝟐
𝟒 𝟒

Solution 4 by Angel Plaza-Spain
By doing 𝟑√𝒙𝒚 = 𝒂, 𝟑√𝒚𝒛 = 𝒃, 𝟑√𝒛𝒙 = 𝒄, the problem becomes
If 𝒂, 𝒃, 𝒄 > 0, 𝑎 + 𝑏 + 𝑐 = 3 then:
𝟑 𝒂𝟑 + 𝒄𝟑 𝟑 𝒃𝟑 + 𝒄𝟑 𝟑 𝒄𝟑 + 𝒃𝟑
√ +√ +√ ≥𝟑
𝟐 𝟐 𝟐
By power mean arithmetic mean inequality:
𝟑 𝒂𝟑 + 𝒄𝟑 𝟑 𝒃𝟑 + 𝒄𝟑 𝟑 𝒄𝟑 + 𝒃𝟑 𝒂 + 𝒄 𝒃 + 𝒄 𝒄 + 𝒂
√ +√ +√ ≥ + +
𝟐 𝟐 𝟐 𝟐 𝟐 𝟐
Equality holds for 𝒂 = 𝒃 = 𝒄 = 𝟏 ⇔ 𝒙 = 𝒚 = 𝒛 = 𝟏

JP.401 Find all sets 𝒙, 𝒚, 𝒛 of positive integers such that
𝟏 𝟏 𝟏 𝟏
+ + =
𝒙𝒚 𝒚𝒛 𝒛𝒙 𝟐
𝟏 𝟏 𝟏 𝟏
We have + 𝒚𝒛 + 𝒛𝒙 = 𝟐 ⇔ 𝒙𝒚𝒛 = 𝟐(𝒙 + 𝒚 + 𝒛)
𝒙𝒚
Assume WLOG that 𝒙 ≤ 𝒚 ≤ 𝒛. Then it follows that (since 𝒙, 𝒚, 𝒛 are positive integers).

www.ssmrmh.ro
𝒙𝒚𝒛 ≤ 𝟔𝒛 ⇒ 𝒙𝒚 ≤ 𝟔 or 𝒙𝟐 ≤ 𝟔 namely 𝒙 = 𝟏 or 𝒙 = 𝟐.
For 𝒙 = 𝟏 ⇒ 𝒚𝒛 = 𝟐 + 𝟐𝒚 + 𝟐𝒛 ⇔ 𝒚𝒛 − 𝟐𝒚 − 𝟐𝒛 = 𝟐 ⇔ (𝒚 − 𝟐)(𝒛 − 𝟐) = 𝟔 = 𝟐 ⋅ 𝟑 with
𝒚 − 𝟐 being the smaller positive factor.
This leads to the two solutions (𝒙, 𝒚, 𝒛) = (𝟏, 𝟑, 𝟖), (𝟏, 𝟒, 𝟓).
For 𝒙 = 𝟐, we have (𝒚 − 𝟏)(𝒛 − 𝟏) = 𝟑 = 𝟏 ⋅ 𝟑 namely 𝒚 = 𝟐 and 𝒛 = 𝟒.
So, (𝒙, 𝒚, 𝒛) ∈ {(𝟏, 𝟑, 𝟖), (𝟏, 𝟒, 𝟓), (𝟐, 𝟐, 𝟒)}
Solution 2 by Ertan Yildirim-Turkiye
𝟏 𝟏 𝟏 𝒙+𝒚+𝒛 𝟏
+ + = = ⇒ 𝟐(𝒙 + 𝒚 + 𝒛) = 𝒙𝒚𝒛; (∗)
𝒙𝒚 𝒚𝒛 𝒛𝒙 𝒙𝒚𝒛 𝟐
Suppose that: 𝒙 ≥ 𝒚 ≥ 𝒛 ⇒ 𝟐(𝒙 + 𝒚 + 𝒛) ≤ 𝟐(𝒙 + 𝒙 + 𝒙) = 𝟔𝒙
⇒ 𝒙𝒚𝒛 ≤ 𝟔𝒙 ⇒ 𝒚𝒛 ≤ 𝟔
(𝒊)𝒚𝒛 = 𝟔 ⇒ 𝒚 = 𝟔, 𝒛 = 𝟏.
𝟕
(∗): 𝟐(𝒙 + 𝟕) = 𝟔𝒙 ⇒ 𝟏𝟒 = 𝟒𝒙 ⇒ 𝒙 = ∉ℤ
𝟐
(𝒊𝒊)𝒚𝒛 = 𝟔 ⇒ 𝒚 = 𝟑, 𝒛 = 𝟐
𝟓
(∗): 𝟐(𝒙 + 𝟓) = 𝟔𝒙 ⇒ 𝟏𝟎 = 𝟒𝒙 ⇒ 𝒙 = ∉ℤ
𝟐
(𝒊𝒊𝒊)𝒚𝒛 = 𝟓 ⇒ 𝒚 = 𝟓, 𝒛 = 𝟏
(∗): 𝟐(𝒙 + 𝟔) = 𝟓𝒙 ⇒ 𝟏𝟐 = 𝟑𝒙 ⇒ 𝒙 = 𝟒
⇒ (𝒙, 𝒚, 𝒛) ∈ {(𝟒, 𝟓, 𝟏), (𝟒, 𝟏, 𝟓), (𝟓, 𝟏, 𝟒), (𝟓, 𝟒, 𝟏), (𝟏, 𝟒, 𝟓), (𝟏, 𝟓, 𝟒)}
(𝒊𝒗)𝒚𝒛 = 𝟒, 𝒚 = 𝟐, 𝒛 = 𝟐
(∗): 𝟐(𝒙 + 𝟒) = 𝟒𝒙 ⇒ 𝟖 = 𝟐𝒙 ⇒ 𝒙 = 𝟒
(𝒙, 𝒚, 𝒛) ∈ {(𝟒, 𝟐, 𝟐), (𝟐, 𝟒, 𝟐), (𝟐, 𝟐, 𝟒)}
(𝒗)𝒚𝒛 = 𝟒, 𝒚 = 𝟒, 𝒛 = 𝟏
(∗): 𝟐(𝒙 + 𝟓) = 𝟒𝒙 ⇒ 𝒙 = 𝟓 (same case in (𝒊𝒊𝒊))
(𝒗𝒊)𝒚𝒛 = 𝟑, 𝒚 = 𝟑, 𝒛 = 𝟏
(∗): 𝟐(𝒙 + 𝟒) = 𝟑𝒙 ⇒ 𝒙 = 𝟖
(𝒙, 𝒚, 𝒛) ∈ {(𝟖, 𝟑, 𝟏), (𝟖, 𝟏, 𝟑), (𝟏, 𝟑, 𝟖), (𝟏, 𝟖, 𝟑), (𝟑, 𝟏, 𝟖), (𝟑, 𝟖, 𝟏)}
(𝒗𝒊𝒊)𝒚𝒛 = 𝟐, 𝒚 = 𝟐, 𝒛 = 𝟏
(∗): 𝟐(𝒙 + 𝟑) = 𝟐𝒙 ⇒ 𝟔 = 𝟎 no solution!

www.ssmrmh.ro
(𝒗𝒊𝒊𝒊)𝒚𝒛 = 𝟏, 𝒚 = 𝒛 = 𝟏
(∗): 𝟐(𝒙 + 𝟐) = 𝒙 ⇒ 𝒙 = −𝟒 ∉ ℤ+
𝟏 𝟏 𝟏 𝟏 𝒙+𝒚+𝒛 𝟏
𝒙, 𝒚, 𝒛 ∈ ℕ∗ , + + = ⇒ = ⇔ 𝟐(𝒙 + 𝒚 + 𝒛) = 𝒙𝒚𝒛; (𝟏)
𝒙𝒚 𝒚𝒛 𝒛𝒙 𝟐 𝒙𝒚𝒛 𝟐
Let suppose: 𝒙 ≤ 𝒚 ≤ 𝒛 ⇒ 𝟐(𝒙 + 𝒚 + 𝒛) ≤ 𝟔𝒛
𝒙 = 𝟏, 𝒚 = 𝟏; (𝟏) ⇒ 𝒛 = 𝟐𝒛 + 𝟒 impossible.
𝒙 = 𝟏, 𝒚 = 𝟐; (𝟏) ⇒ 𝟐𝒛 = 𝟐(𝟑 + 𝒛) impossible.
𝒙 = 𝟏, 𝒚 = 𝟑; (𝟏) ⇒ 𝒛 = 𝟖
𝒙 = 𝟏, 𝒚 = 𝟒; (𝟏) ⇒ 𝒛 = 𝟓
𝒙 = 𝟏, 𝒚 = 𝟒; (𝟏) ⇒ 𝒛 = 𝟒
𝒙 = 𝟏, 𝒚 = 𝟓; (𝟏) ⇒impossible.
𝒙 = 𝟐, 𝒚 = 𝟏 impossible.
𝒙 = 𝟐, 𝒚 = 𝟐; (𝟏) ⇒ 𝒛 = 𝟒
𝒙 = 𝟐, 𝒚 = 𝟑 impossible.
𝒙 = 𝟑, 𝒚 = 𝟏; (𝟏) ⇒ 𝒛 = 𝟖
𝒙 = 𝟑, 𝒚 = 𝟐 impossible.
𝒙 = 𝟒, 𝒚 = 𝟏; (𝟏) ⇒ 𝒛 = 𝟓
𝒙 = 𝟓, 𝒚 = 𝟏; (𝟏) ⇒ 𝒛 = 𝟒
𝒙 = 𝟔, 𝒚 = 𝟏; (𝟏) ⇒impossible.
Hence,
(𝒙, 𝒚, 𝒛) ∈ {(𝟏, 𝟑, 𝟖), (𝟏, 𝟒, 𝟓), (𝟏, 𝟓, 𝟒), (𝟐, 𝟐, 𝟒), (𝟑, 𝟏, 𝟖), (𝟒, 𝟏, 𝟓), (𝟓, 𝟏, 𝟒)}
By applying permutations 𝒙 ≤ 𝒚 ≤ 𝒛 and 𝒙 ≤ 𝒛 ≤ 𝒚, we get:
(𝟏, 𝟑, 𝟖), (𝟏, 𝟒, 𝟓), (𝟏, 𝟓, 𝟒), (𝟐, 𝟐, 𝟒), (𝟑, 𝟏, 𝟖), (𝟒, 𝟏, 𝟓), (𝟓, 𝟏, 𝟒),
(𝒙, 𝒚, 𝒛) ∈ { (𝟏, 𝟒, 𝟓), (𝟏, 𝟓, 𝟒), (𝟒, 𝟏, 𝟓), (𝟒, 𝟓, 𝟏), (𝟓, 𝟏, 𝟒), (𝟓, 𝟒, 𝟏) }
(𝟐, 𝟐, 𝟒), (𝟐, 𝟒, 𝟐), (𝟒, 𝟐, 𝟐)
JP.402 In acute 𝚫𝑨𝑩𝑪 the following relationship holds:

𝑹 − 𝟐𝒓 ≤ 𝟑(𝐦𝐚𝐱{𝒉𝒂 , 𝒉𝒃, 𝒉𝒄 } − 𝐦𝐢𝐧{𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 })
Proposed by Cristian Miu-Romania
Let us recall two inequalities. The first one is Bankhoff inquality:
∑𝒉𝒂 ≤ 𝟐𝑹 + 𝟓𝒓 and the second one is Erdos inequality. In any acute triangle,
𝑹 + 𝒓 ≤ 𝐦𝐚𝐱{𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 }. Now, we can write:
𝟐𝑹 + 𝟓𝒓
𝐦𝐢𝐧{𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 } ≤ ≤ 𝑹 + 𝒓 ≤ 𝐦𝐚𝐱{𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 }
𝟑

www.ssmrmh.ro
𝟐𝒓+𝟓𝒓 𝑹−𝟐𝒓
So, we obtain: 𝐦𝐚𝐱{𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 } − 𝐦𝐢𝐧{𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 } ≥ 𝑹 + 𝒓 − 𝟑 = 𝟑
𝟏 𝒔𝟐 +𝒓𝟐 +𝟒𝑹𝒓
Bankhoff inequality is easy to prove, because ∑𝒉𝒂 = 𝟐𝑹 ∑𝒂𝒃 = .
𝟐𝑹
𝒔𝟐 +𝒓𝟐 +𝟒𝑹𝒓
But ≤ 𝟐𝑹 + 𝟓𝒓. This inequality can be proved using Gerretsen inequality
𝟐𝑹
𝒔𝟐 ≤ 𝟒𝑹𝟐 + 𝟒𝑹𝒓 + 𝟑𝒓𝟐 .

Let us prove Erdos inequality. It is easy to see that ∑𝒂 𝐜𝐨𝐭 𝑨 = 𝟐(𝑹 + 𝒓) and
∑𝒂𝟐 𝐜𝐨𝐭 𝑨 = 𝟒𝑭, where 𝑭 −is area of triangle 𝑨𝑩𝑪. So, we obtain:
𝐦𝐢𝐧{𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 } ≤ 𝑹 + 𝒓 ≤ 𝐦𝐚𝐱{𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 } because 𝒙, 𝒚, 𝒛 are real numbers and 𝒖, 𝒗, 𝒘
𝒙 𝒚 𝒛 𝒙+𝒚+𝒛 𝒙 𝒚 𝒛
are positive, then 𝐦𝐢𝐧 {𝒖 , 𝒗 , 𝒘} ≤ 𝒖+𝒗+𝒘 ≤ 𝐦𝐚𝐱 {𝒖 , 𝒗 , 𝒘}

𝑾𝑳𝑶𝑮, 𝒘𝒆 𝒎𝒂𝒚 𝒂𝒔𝒔𝒖𝒎𝒆 𝒕𝒉𝒂𝒕 ∶ 𝒂 ≥ 𝒃 ≥ 𝒄 → 𝒉𝒂 ≤ 𝒉𝒃 ≤ 𝒉𝒄 → (∗) ↔ 𝑹 − 𝟐𝒓
𝟔𝒔𝒓(𝒂 − 𝒄)
≤ 𝟑( 𝒉 𝒄 − 𝒉 𝒂 ) =
𝒄𝒂
𝟔𝒓(𝒂 − 𝒄) 𝑹 𝟔𝒂 − 𝟒𝒄
𝑺𝒊𝒏𝒄𝒆 𝒔 > 𝑎 → 𝐼𝒕′ 𝒔 𝒔𝒖𝒇𝒇𝒊𝒄𝒆𝒔 𝒕𝒐 𝒑𝒓𝒐𝒗𝒆 𝒕𝒉𝒂𝒕 ∶ 𝑹 − 𝟐𝒓 ≤ ↔ ≤
𝒄 𝒓 𝒄
𝟔𝒂 − 𝟒𝒄 𝟐𝒂 + 𝟒(𝒂 − 𝒄) 𝒂≥𝒄 𝟐𝒂 ? 𝑹
𝑾𝒆 𝒉𝒂𝒗𝒆 ∶ = ⏞
≥ ⏞
≥
𝒄 𝒄 𝒄 𝒓
𝟐𝒂𝒃𝒄
=
(−𝒂 + 𝒃 + 𝒄)(𝒂 − 𝒃 + 𝒄)(𝒂 + 𝒃 − 𝒄)
(−𝒂 + 𝒃 + 𝒄)(𝒂 − 𝒃 + 𝒄)(𝒂 + 𝒃 − 𝒄) ≥ 𝒃𝒄𝟐 ↔ (−𝒂 + 𝒃 + 𝒄)(𝒂𝟐 − 𝒃𝟐 − 𝒄𝟐 + 𝟐𝒃𝒄)

≥ 𝒃𝒄𝟐
↔ (−𝒂 + 𝒃)(−𝟐𝒃𝒄. 𝒄𝒐𝒔 𝑨 − 𝒄(𝒂 − 𝒃) + 𝒄(𝒂 + 𝒃)) + 𝒄(𝒂𝟐 − 𝒃𝟐 − 𝒄𝟐 + 𝟐𝒃𝒄) ≥ 𝒃𝒄𝟐
↔ (𝒂 − 𝒃)[𝟐𝒃𝒄. 𝒄𝒐𝒔 𝑨 + 𝒄(𝒂 − 𝒃)] + 𝒄(𝒃𝟐 − 𝒂𝟐 ) + 𝒄(𝒂𝟐 − 𝒃𝟐 − 𝒄𝟐 + 𝟐𝒃𝒄) ≥ 𝒃𝒄𝟐
↔ (𝒂 − 𝒃)[𝟐𝒃𝒄. 𝒄𝒐𝒔 𝑨 + 𝒄(𝒂 − 𝒃)] + 𝒄𝟐 (𝒃 − 𝒄)

≥ 𝟎 𝒘𝒉𝒊𝒄𝒉 𝒊𝒔 𝒕𝒓𝒖𝒆 (∴ 𝒂 ≥ 𝒃 ≥ 𝒄, 𝒄𝒐𝒔 𝑨 ≥ 𝟎)
𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆, 𝑹 − 𝟐𝒓 ≤ 𝟑(𝒎𝒂𝒙{𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 } − 𝒎𝒊𝒏{𝒉𝒂 , 𝒉𝒃 , 𝒉𝒄 }).
JP.403 If 𝒙, 𝒚, 𝒛 ∈ (𝟎, 𝟏) then in 𝚫𝑨𝑩𝑪 the following relationship holds:

𝟏 𝟏 𝟏 𝟑√𝟑
+ + ≥
(𝒚 + 𝒛)(𝟏 − 𝒙𝟐 )𝒉𝟒𝒂 (𝒛 + 𝒙)(𝟏 − 𝒚𝟐 )𝒉𝟒𝒃 (𝒙 + 𝒚)(𝟏 − 𝒛𝟐 )𝒉𝟒𝒄 𝟒𝑭𝟐
Proposed by D.M. Bătineţu-Giurgiu, Daniel Sitaru-Romania

www.ssmrmh.ro
𝟑
𝟐( 𝟐 )𝟐
𝑨𝑴−𝑮𝑴 𝟐𝒙𝟐 + 𝟏 − 𝒙𝟐 + 𝟏 − 𝒙𝟐 𝟖
𝟐𝒙 𝟏 − 𝒙 ≤ ( ) =
𝟑 𝟐𝟕
𝟒 𝟐
𝒙𝟐 ( 𝟏 − 𝒙𝟐 ) 𝟐 ≤ ⇒ 𝒙( 𝟏 − 𝒙𝟐 ) ≤ ; ( 𝟏)
𝟐𝟕 𝟑√𝟑
𝟏 𝒙 (𝟏) 𝟑√𝟑 𝒙
∑ 𝟐 𝟒
= ∑ 𝟐 𝟒
≥ ∑ =
(𝒚 + 𝒛)(𝟏 − 𝒙 )𝒉𝒂 𝒙(𝒚 + 𝒛)(𝟏 − 𝒙 )𝒉𝒂 𝟐 (𝒚 + 𝒛)𝒉𝟒𝒂
𝟑√𝟑 𝒙𝒂𝟒 𝟑√𝟑 𝟏 𝒙 𝑻𝒔𝒊𝒏𝒕𝒔𝒊𝒇𝒂𝒔

= ∑ 𝟒
= ⋅ 𝟒
∑ 𝒂𝟒 ≥
𝟐 (𝒚 + 𝒛)(𝒂𝒉𝒂 ) 𝟐 𝟏𝟔𝑭 𝒚+𝒛
𝟑√𝟑 𝟐
𝟑√𝟑
≥ ⋅ 𝟖𝑭 =
𝟑𝟐𝑭𝟒 𝟒𝑭𝟐
√𝟑
Equality holds for 𝒂 = 𝒃 = 𝒄; 𝒙 = 𝒚 = 𝒛 = .
𝟑
𝟑
√𝟑 √𝟑 √𝟑 √𝟑
𝑩𝒚 𝑨𝑴 − 𝑮𝑴, 𝒘𝒆 𝒉𝒂𝒗𝒆 ∶ 𝒙𝟐 + + ≥ 𝟑 √ 𝒙𝟐 . . =𝟏 →
𝟗𝒙 𝟗𝒙 𝟗𝒙 𝟗𝒙
𝟐√𝟑
𝟏 − 𝒙𝟐 ≤ (𝑨𝒏𝒅 𝒂𝒏𝒂𝒍𝒐𝒈𝒔)
𝟗𝒙
𝟗 𝒙 𝒂 𝟒 𝟑√𝟑 ( 𝒙 + 𝒚 + 𝒛) − ( 𝒚 + 𝒛) 𝟒
𝑳𝑯𝑺(∗) ≥ ∑ .
.( ) = ∑ .𝒂
𝟐√𝟑 𝒚 + 𝒛 𝟐𝑭 𝟑𝟐𝑭𝟒 𝒚+𝒛
𝟑√𝟑 𝒂𝟒
= [(∑ 𝒙 ) (∑ ) − ∑ 𝒂𝟒 ] ≥
𝟑𝟐𝑭𝟒 𝒚+𝒛
𝟐
𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎
𝟑√𝟑 (𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 )𝟐 𝟑√𝟑
⏞
≥ [(∑ 𝒙 ) . − ∑ 𝒂𝟒
] = [(∑ 𝒂𝟐 ) − 𝟐 ∑ 𝒂𝟒 ]
𝟑𝟐𝑭𝟒 𝟐 ( 𝒙 + 𝒚 + 𝒛) 𝟔𝟒𝑭𝟒
𝟑√𝟑
= (𝟐 ∑ 𝒂𝟐 𝒃𝟐 − ∑ 𝒂𝟒 )
𝟔𝟒𝑭𝟒

www.ssmrmh.ro
𝟏 𝟑√𝟑 𝟑√𝟑
𝑺𝒊𝒏𝒄𝒆 𝟐 ∑ 𝒂𝟐 𝒃𝟐 − ∑ 𝒂𝟒 = 𝟏𝟔𝑭𝟐 → ∑ 𝟒 ≥ 𝟒
. 𝟏𝟔𝑭𝟐 = .
(𝒚 + 𝒛)(𝟏 − 𝒙𝟐 )𝒉𝒂 𝟔𝟒𝑭 𝟒𝑭𝟐
𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆,
𝟏 𝟏 𝟏 𝟑√𝟑
+ + ≥ .
(𝒚 + 𝒛)(𝟏 − 𝒙𝟐 ) 𝒉 𝒂 𝟒 (𝒛 + 𝒙)(𝟏 − 𝒚𝟐 ) 𝒉 𝒃 𝟒 (𝒙 + 𝒚)(𝟏 − 𝒛𝟐 ) 𝒉 𝒄 𝟒 𝟒𝑭𝟐
√𝟑
𝟑
Solution 3 by Soumitra Mandal-Chandar Nagore-India

𝟐𝑭 𝟐𝑭 𝟐𝑭
𝟐 ∑ 𝒂𝟐 𝒃𝟐 − ∑ 𝒂𝟒 = 𝟏𝟔𝑭𝟐 , 𝒉𝒂 = , 𝒉𝒃 = , 𝒉𝒄 =
𝒂 𝒃 𝒄
Let 𝒇(𝒙) = 𝒙 − 𝒙𝟑 , ∀𝒙 ≥ 𝟎, 𝒇′ (𝒙) = 𝟏 − 𝟑𝒙𝟐 , 𝒇′′ (𝒙) = −𝟔𝒙

𝟏 𝟏 𝟔 𝟏
𝒇′ (𝒙) = 𝟎 ⇒ 𝒙 = ⇒ 𝒇′′ ( ) = − < 0 ⇒ 𝐦𝐚𝐱 𝒇(𝒙) is attains at 𝒙 =
√𝟑 √𝟑 √𝟑 √𝟑
𝟏 𝟐
𝒇( ) ≥ 𝒇 ( 𝒙) ⇒ ≥ 𝒇(𝒙)
√𝟑 𝟑√𝟑
𝟏 𝒙 𝟏 𝟏 𝟑√𝟑 𝒙 𝟏
∑ 𝟐 𝟒
=∑ ⋅ 𝟑
⋅ 𝟒≥ ∑ ⋅ 𝟒=
(𝒚 + 𝒛)(𝟏 − 𝒙 )𝒉𝒂 𝒚 + 𝒛 𝒙 − 𝒙 𝒉𝒂 𝟐 𝒚 + 𝒛 𝒉𝒂
𝟑√𝟑 𝒙 𝟏
= ∑( + 𝟏 − 𝟏) ⋅ 𝟒 =
𝟐 𝒚+𝒛 𝒉𝒂
𝒄𝒚𝒄
𝟏 𝟐
𝟑√𝟑 ) 𝟑√𝟑 𝟏 ( 𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎
𝒉𝟐𝒂
= ( )
𝒙+𝒚+𝒛 ∑ − ∑ 𝟒 ≥
𝟐 𝒚+𝒛 𝟐 𝒉𝒂
𝟐
𝟐
𝟑√𝟑 𝟏 𝟑√𝟑 𝟏 𝟑√𝟑 𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 𝟑√𝟑 𝒂𝟒
≥ (∑ 𝟐 ) − ∑ 𝟒= ( ) − ∑ =
𝟒 𝒉𝒂 𝟐 𝒉𝒂 𝟒 𝟒𝑭𝟐 𝟐 𝟏𝟔𝑭𝟐
𝟑√𝟑 (𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 )𝟐 𝟐(𝒂𝟒 + 𝒃𝟒 + 𝒄𝟒 )

= { − }=
𝟒𝑭𝟐 𝟏𝟔𝑭𝟐 𝟏𝟔𝑭𝟐
𝟑√𝟑 𝟐(𝒂𝟐 𝒃𝟐 + 𝒃𝟐 𝒄𝟐 + 𝒄𝟐 𝒂𝟐 ) − 𝒂𝟒 − 𝒃𝟒 − 𝒄𝟒 𝟑√𝟑
= ⋅ =
𝟒𝑭𝟐 𝟏𝟔𝑭𝟐 𝟒𝑭𝟐
√𝟑
𝟑

www.ssmrmh.ro
𝝅 𝟏 𝟏
JP.404 If 𝒙, 𝒚 > 0; 𝑧 ∈ [𝟎, ] ; + ≥ 𝟏 then:
𝟐 𝒙+𝐬𝐢𝐧 𝒛 𝒚+𝐬𝐢𝐧 𝒛
𝟏 𝟏
+ ≥ 𝟐 𝐬𝐢𝐧 𝒛
𝒙 𝒚
Proposed by D.M.Bătineţu-Giurgiu, Daniel Sitaru-Romania
𝟏 𝑨𝑴−𝑮𝑴 𝐬𝐢𝐧 𝒛
(𝒙 + 𝐬𝐢𝐧 𝒛) ( + 𝐬𝐢𝐧 𝒛) ≥ 𝟐√𝒙 ⋅ 𝐬𝐢𝐧 𝒛 ⋅ 𝟐√ = 𝟒 𝐬𝐢𝐧 𝒛
𝒙 𝒙
𝟏 𝟒 𝐬𝐢𝐧 𝒛
+ 𝐬𝐢𝐧 𝒛 ≥ ; ( 𝟏)
𝒙 𝒙 + 𝐬𝐢𝐧 𝒛
Analogous,
𝟏 𝟒 𝐬𝐢𝐧 𝒛
+ 𝐬𝐢𝐧 𝒛 ≥ ; ( 𝟐)
𝒚 𝒚 + 𝐬𝐢𝐧 𝒛
By adding (1),(2) we get:
𝟏 𝟏 𝟒 𝐬𝐢𝐧 𝒛 𝟒 𝐬𝐢𝐧 𝒛
+ + 𝟐 𝐬𝐢𝐧 𝒛 ≥ + ⇔
𝒙 𝒚 𝒙 + 𝐬𝐢𝐧 𝒛 𝒚 + 𝐬𝐢𝐧 𝒛
𝟏 𝟏 𝟒 𝐬𝐢𝐧 𝒛 𝟒 𝐬𝐢𝐧 𝒛 𝟏 𝟏
+ ≥ + − 𝟐 𝐬𝐢𝐧 𝒛 = 𝟐 𝐬𝐢𝐧 𝒛 (𝟐 ( + ) − 𝟏) ≥
𝒙 𝒚 𝒙 + 𝐬𝐢𝐧 𝒛 𝒚 + 𝐬𝐢𝐧 𝒛 𝒙 + 𝐬𝐢𝐧 𝒛 𝒚 + 𝐬𝐢𝐧 𝒛
≥ 𝟐 𝐬𝐢𝐧 𝒛 (𝟐 ⋅ 𝟏 − 𝟏) = 𝟐 𝐬𝐢𝐧 𝒛
𝝅
Equality holds for 𝒙 = 𝒚 = 𝟏; 𝒛 = .
𝟐

𝟏 𝟏
If 𝒛 = 𝟎 →we need to prove that∶ + 𝒚 ≥ 𝟎 which is true.
𝒙
𝑨𝑴−𝑮𝑴
𝟏 𝟏 𝟏 𝟏
𝐈𝐟 𝒛 ≠ 𝟎 → 𝐬𝐢𝐧 𝒛 > 0 → 1 ≤ + ⏞
≤ + ≤
𝒙 + 𝐬𝐢𝐧 𝒛 𝒚 + 𝐬𝐢𝐧 𝒛 𝟐√𝒙. 𝐬𝐢𝐧 𝒛 𝟐√𝒚. 𝐬𝐢𝐧 𝒛
𝑪𝑩𝑺
𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
⏞
≤ . √( + )( + ) → 𝟏 ≤ ( + ).
𝟐 𝐬𝐢𝐧 𝒛 𝐬𝐢𝐧 𝒛 𝒙 𝒚 𝟐 𝐬𝐢𝐧 𝒛 𝒙 𝒚
Therefore,
𝟏 𝟏
+ ≥ 𝟐 𝐬𝐢𝐧 𝒛.
𝒙 𝒚

www.ssmrmh.ro
𝝅
Equality holds for 𝒙 = 𝒚 = 𝟏; 𝒛 = 𝟐 .

If 𝒛 = 𝟎, then is nothing to prove.
𝝅
Assume 𝟎 < 𝑧 ≤ ⇔ 𝐬𝐢𝐧 𝒛 > 0
𝟐
𝟏 𝟏 𝟏 𝟏
𝟏≤ + ≤ +
𝒙 + 𝐬𝐢𝐧 𝒛 𝒚 + 𝐬𝐢𝐧 𝒛 𝟐√𝒙 ⋅ 𝐬𝐢𝐧 𝒛 𝟐√𝒚 ⋅ 𝐬𝐢𝐧 𝒛
𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
⇒ 𝟐√𝐬𝐢𝐧 𝒛 ≤ + ⇒ 𝟒 𝐬𝐢𝐧 𝒛 ≤ + + ≤ + + +
√𝒙 √𝒚 𝒙 𝒚 √𝒙𝒚 𝒙 𝒚 𝒙 𝒚
Hence,
𝟏 𝟏
𝟐 𝐬𝐢𝐧 𝒛 ≤ +
𝒙 𝒚
𝝅
Solution 4 by Amrit Awasthi-India

𝝅
𝟎≤𝒛≤ ⇒ 𝟎 ≤ 𝐬𝐢𝐧 𝒛 ≤ 𝟏
𝟐
𝟏 𝟏
𝒙 + 𝐬𝐢𝐧 𝒛 ≤ (𝟏 + 𝒙) ⇒ ≥
𝒙 + 𝐬𝐢𝐧 𝒛 𝟏 + 𝒙
Similarly,
𝟏 𝟏
≥
𝒚 + 𝐬𝐢𝐧 𝒛 𝟏 + 𝒚
Adding, we get:
𝟏 𝟏 𝟏 𝟏 𝒙+𝒚+𝟐
+ ≥ + =
𝒙 + 𝐬𝐢𝐧 𝒛 𝒚 + 𝐬𝐢𝐧 𝒛 𝟏 + 𝒙 𝟏 + 𝒚 (𝟏 + 𝒙)(𝟏 + 𝒚)
Now it’s given that
𝟏 𝟏
+ ≥𝟏
𝒙 + 𝐬𝐢𝐧 𝒛 𝒚 + 𝐬𝐢𝐧 𝒛
Hence, we must have
( 𝒙 + 𝒚 + 𝟐)
≥ 𝟏 ⇔ 𝒙 + 𝒚 + 𝟐 ≥ 𝟏 + 𝒙 + 𝒚 + 𝒙𝒚 ⇒ 𝒙𝒚 ≤ 𝟏
(𝟏 + 𝒙)(𝟏 + 𝒚)
𝟏
Now, as 𝒙𝒚 ≤ 𝟏 ⇒ √𝒙𝒚 ≤ 𝟏 ⇒ ≥ 𝟏; (∗)
√𝒙𝒚

www.ssmrmh.ro
Therefore,
𝟏 𝟏 𝑨𝑴−𝑮𝑴 𝟐 (∗) (∗∗)
+ ≥ ≥ 𝟐 ≥ 𝟐 𝐬𝐢𝐧 𝒛
𝒙 𝒚 √𝒙𝒚
𝟏 𝟏
(∗∗) ⇔ 𝟏 ≥ 𝐬𝐢𝐧 𝒛 true. So, + ≥ 𝟐 𝐬𝐢𝐧 𝒛
𝒙 𝒚
𝝅

𝒂𝟑 𝒃𝟑 𝒄𝟑 𝟒√𝟑𝑭
+ + ≥
𝟐𝒃 + 𝟐𝒄 − 𝒂 𝟐𝒄 + 𝟐𝒂 − 𝒃 𝟐𝒂 + 𝟐𝒃 − 𝒄 𝟑
Proposed by D.M. Bătineţu-Giurgiu, Neculai Stanciu-Romania
𝒂𝟑 𝒂𝟒 𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎
∑ =∑ ≥
𝟐(𝒃 + 𝒄) − 𝒂 𝟐𝒂𝒃 + 𝟐𝒂𝒄 − 𝒂𝟐
(𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 )𝟐 (𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 )𝟐
≥ ≥ =
𝟒(𝒂𝒃 + 𝒃𝒄 + 𝒄𝒂) − (𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ) 𝟒(𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ) − (𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 )
𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 𝑰𝒐𝒏𝒆𝒔𝒄𝒖−𝑾𝒆𝒊𝒕𝒛𝒆𝒏𝒃𝒐𝒄𝒌 𝟒√𝟑𝑭
= ≥
𝟑 𝟑
𝒂𝟑 𝒂𝟒 𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎 (𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 )𝟐
∑ =∑ ≥ ; ( 𝟏)
𝟐(𝒃 + 𝒄) − 𝒂 𝟐𝒂(𝒃 + 𝒄) − 𝒂𝟐 𝟒(𝒂𝒃 + 𝒃𝒄 + 𝒄𝒂) − 𝒂𝟐 − 𝒃𝟐 − 𝒄𝟐
We use 𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ≥ 𝒂𝒃 + 𝒃𝒄 + 𝒄𝒂 for both counter and denominator and we have:

𝒂𝟑 (𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 )𝟐
∑ ≥ ≥
𝟐(𝒃 + 𝒄) − 𝒂 𝟒(𝒂𝒃 + 𝒃𝒄 + 𝒄𝒂) − 𝒂𝟐 − 𝒃𝟐 − 𝒄𝟐
𝒄𝒚𝒄
(𝒂𝒃 + 𝒃𝒄 + 𝒄𝒂)𝟐 𝒂𝒃 + 𝒃𝒄 + 𝒄𝒂 𝟒√𝟑𝑭

≥ = ≥
𝟑(𝒂𝒃 + 𝒃𝒄 + 𝒄𝒂) 𝟑 𝟑

www.ssmrmh.ro
PROBLEMS FOR SENIORS
SP.391 Find:
∑𝒌𝒊=𝟏 𝒊𝟐
−𝟏
𝒏 𝒌
𝛀 = 𝐥𝐢𝐦 ∑ (𝟏 + (∑ 𝒊𝟐 ) ) − 𝒏𝒆
𝒏→∞
𝒌=𝟏 𝒊=𝟏
( )
Proposed by Florică Anastase-Romania
𝟏 𝟏 ′ 𝒙
( )
𝐥𝐢𝐦
( 𝟏 + 𝒙) 𝒙 − 𝒆 𝑳′ 𝑯
= 𝐥𝐢𝐦 (𝟏 + ) (
𝟏 𝒙 𝐥𝐨𝐠(𝟏 + 𝒙)
) = 𝒆 ⋅ 𝐥𝐢𝐦 𝒙 + 𝟏 − 𝐥𝐨𝐠 𝒙 + 𝟏 =
𝒙→𝟎 𝒙 𝒙→𝟎 𝒙 𝒙 𝒙→𝟎 𝒙𝟐
𝒙 − (𝒙 + 𝟏) 𝐥𝐨𝐠(𝒙 + 𝟏) 𝑳′ 𝑯 𝟏 − 𝐥𝐨𝐠(𝒙 + 𝟏) − 𝟏
= 𝒆 ⋅ 𝐥𝐢𝐦 = 𝒆 ⋅ 𝐥𝐢𝐦
𝒙→𝟎 𝒙 𝟐 ( 𝒙 + 𝟏) 𝒙→𝟎 𝟑𝒙𝟐 + 𝟐𝒙
−𝟏 𝒆
= 𝒆 ⋅ 𝐥𝐢𝐦 =−
𝒙→𝟎 (𝒙 + 𝟏)(𝟔𝒙 + 𝟐) 𝟐
𝟏
( 𝟏 + 𝒙) 𝒙 − 𝒆 𝒆
𝐁𝐞𝐜𝐚𝐮𝐬𝐞: 𝐥𝐢𝐦 = − 𝐭𝐡𝐞𝐧 ∀𝜺 > 0, ∃𝛿 (𝜺) > 0 𝐬𝐮𝐜𝐡 𝐭𝐡𝐚𝐭 |𝐱| < 𝜹(𝛆)
𝒙→𝟎 𝒙 𝟐
𝟏 𝟏
(𝟏 + 𝒙)𝒙 − 𝒆 𝒆 ( 𝟏 + 𝒙) 𝒙 − 𝒆 𝒆
⇒| − ( )| < 𝜀 ⇔ −𝜀 < − (− ) < 𝜀
𝒙 𝟐 𝒙 𝟐
𝟏
𝒆 (𝟏 + 𝒙)𝒙 − 𝒆 𝒆
⇔ −𝜺 − < <𝜀−
𝟐 𝒙 𝟐
𝟏 𝟏 𝟏
𝟎< 𝟐 < ⋯ < < < 𝛿(𝜀)
𝟏 + 𝟐𝟐 + ⋯ + 𝒏𝟐 𝟏𝟐 + 𝟐𝟐 𝟏𝟐
We have the following relationship:
𝟐 𝟏 +𝟐 𝟐 𝟐
𝟏 𝟏 𝟏
𝒆 (𝟏 + ) − 𝒆 𝒆 𝒆 (𝟏 + ) −𝒆 𝒆
−𝜺 − < 𝟏𝟐 < 𝜀 − , −𝜺 − < 𝟏 𝟐 + 𝟐𝟐 <𝜀−
𝟐 𝟏 𝟐 𝟐 𝟏 𝟐
𝟏𝟐 𝟏𝟐 + 𝟐𝟐
⋮
𝟏𝟐+𝟐𝟐 +⋯+𝒏𝟐
𝟏
𝒆 (𝟏 + 𝟏𝟐 + 𝟐𝟐 + ⋯ + 𝒏𝟐 ) −𝒆 𝒆
−𝜺 − < <𝜀−
𝟐 𝟏 𝟐
𝟐 𝟐
𝟏 +𝟐 + ⋯+ 𝒏 𝟐
Summing these up relations, it follows that:

www.ssmrmh.ro
−𝟏 ∑𝒌
𝒊=𝟏 𝒊
𝟐
𝒏 𝒏 𝒌
𝒆 𝟏
(−𝜺 − ) ∑ 𝟐 < ∑ (𝟏 + (∑ 𝒊𝟐 ) ) − 𝒏𝒆 <
𝟐 𝟏 + 𝟐𝟐 + ⋯ + 𝒌 𝟐
𝒌=𝟏 𝒌=𝟏 𝒊=𝟏
𝒏
𝒆 𝟏
< (𝜺 − ) ∑ 𝟐
𝟐 𝟏 + 𝟐 + ⋯ + 𝒌𝟐
𝟐
𝒌=𝟏
−𝟏 ∑𝒌
𝒊=𝟏 𝒊
𝟐
𝒏 𝒌
𝒆 𝒆
⇔ (−𝜺 − ) ⋅ 𝒂𝒏 < ∑ (𝟏 + (∑ 𝒊𝟐 ) ) − 𝒏𝒆 < (𝜺 − ) ⋅ 𝒂𝒏
𝟐 𝟐
𝒌=𝟏 𝒊=𝟏
𝟏 𝟏 𝟏
𝐖𝐡𝐞𝐫𝐞, 𝟐
𝒂𝒏 =
+ 𝟐 𝟐
+ ⋯+ 𝟐 ; 𝒏 ≥ 𝟏.
𝟏 𝟏 +𝟐 𝟏 + 𝟐 + ⋯ + 𝒏𝟐
𝟐
Now, we get:
𝒏 𝒏
𝟔 𝟏 𝟏 𝟒
𝒂𝒏 = ∑ = 𝟔∑( + − )=
𝒌(𝒌 + 𝟏)(𝟐𝒌 + 𝟏) 𝒌 𝒌 + 𝟏 𝟐𝒌 + 𝟏
𝒌=𝟏 𝒌=𝟏
𝒏 𝒏
𝟏 𝟏 𝟏 𝟏
= 𝟔∑( − ) − 𝟐𝟒 ∑ ( − )=
𝒌+𝟏 𝒌 𝟐𝒌 + 𝟏 𝟐𝒌
𝒌=𝟏 𝒌=𝟏
𝟏
= 𝟔( − 𝟏) − 𝟐𝟒(𝑯𝟐𝒏+𝟏 − 𝑯𝟐𝒏 − 𝟏).
𝒏+𝟏
𝐥𝐢𝐦 (𝑯𝒏 − 𝐥𝐨𝐠 𝒏) = 𝜸
𝒏→∞
𝟐𝒏 + 𝟏
𝑯𝟐𝒏+𝟏 − 𝑯𝟐𝒏 = (𝑯𝟐𝒏+𝟏 − 𝐥𝐨𝐠(𝟐𝒏 + 𝟏)) − (𝑯𝟐𝒏 − 𝐥𝐨𝐠 𝟐𝒏) + 𝐥𝐨𝐠
𝟐𝒏
𝐥𝐢𝐦 (𝑯𝟐𝒏+𝟏 − 𝑯𝟐𝒏 ) = 𝐥𝐨𝐠 𝟐 𝐚𝐧𝐝 𝐥𝐢𝐦 𝒂𝒏 = 𝟔(𝟑 − 𝟒 𝐥𝐨𝐠 𝟐)
𝒏→∞ 𝒏→∞
Therefore,
−𝟏 ∑𝒌
𝒊=𝟏 𝒊
𝟐
𝒏 𝒌
𝒆
𝛀 = 𝐥𝐢𝐦 ∑ (𝟏 + (∑ 𝒊𝟐 ) ) − 𝒏𝒆 = − ⋅ 𝟔(𝟑 − 𝟒 𝐥𝐨𝐠 𝟐) =
𝒏→∞ 𝟐
𝒌=𝟏 𝒊=𝟏
( )
= 𝟑𝒆(𝟒 𝐥𝐨𝐠 𝟐 − 𝟑)
Solution 2 by Adrian Popa-Romania
𝒌
𝒌(𝒌 + 𝟏)(𝟐𝒌 + 𝟏)
∑ 𝒊𝟐 =
𝟔
𝒊=𝟏
𝒏 𝒌(𝒌+𝟏)(𝟐𝒌+𝟏)
𝟔 𝟔
∑ (𝟏 + ) =
𝒌(𝒌 + 𝟏)(𝟐𝒌 + 𝟏)
𝒌=𝟏

www.ssmrmh.ro
𝒏(𝒏+𝟏)(𝟐𝒏+𝟏)
𝟏 𝟓 𝟏 𝟏𝟒 𝟔 𝟔
= (𝟏 + 𝟏)𝟏 + (𝟏 + ) + (𝟏 + ) + ⋯ + (𝟏 + )
𝟓 𝟏𝟒 𝒏(𝒏 + 𝟏)(𝟐𝒏 + 𝟏)
≥
𝑩𝒆𝒓𝒏𝒐𝒖𝒍𝒍𝒊
≥ 𝟐𝒏
𝟏 𝒏
𝒆𝒏 = (𝟏 + 𝒏) increasing, 𝒆𝒏 → 𝒆
𝟏 𝟓 𝟏 𝟏𝟒
(𝟏 + 𝟏 )𝟏 < 𝑒, (𝟏 + ) < 𝑒, (𝟏 + ) < 𝑒, …,
𝟓 𝟏𝟒
𝒏(𝒏+𝟏)(𝟐𝒏+𝟏)
𝟔 𝟔
(𝟏 + ) <𝑒
𝒏(𝒏 + 𝟏)(𝟐𝒏 + 𝟏)
−𝟏 ∑𝒌
𝒊=𝟏 𝒊
𝟐
𝒏 𝒏 𝒌
𝒆 𝟏
(−𝜺 − ) ∑ 𝟐 < ∑ (𝟏 + (∑ 𝒊𝟐 ) ) − 𝒏𝒆 <
𝟐 𝟏 + 𝟐 + ⋯ + 𝒌𝟐
𝟐
𝒌=𝟏 𝒌=𝟏 𝒊=𝟏
𝒏
𝒆 𝟏
< (𝜺 − ) ∑ 𝟐
𝟐 𝟏 + 𝟐𝟐 + ⋯ + 𝒌 𝟐
𝒌=𝟏
−𝟏 ∑𝒌
𝒊=𝟏 𝒊
𝟐
𝒏 𝒌
𝒆 𝒆
⇔ (−𝜺 − ) ⋅ 𝒂𝒏 < ∑ (𝟏 + (∑ 𝒊𝟐 ) ) − 𝒏𝒆 < (𝜺 − ) ⋅ 𝒂𝒏
𝟐 𝟐
𝒌=𝟏 𝒊=𝟏
𝟏 𝟏 𝟏
𝐖𝐡𝐞𝐫𝐞, 𝟐
𝒂𝒏 =
+ 𝟐 𝟐
+ ⋯+ 𝟐 ; 𝒏 ≥ 𝟏.
𝟏 𝟏 +𝟐 𝟏 + 𝟐 + ⋯ + 𝒏𝟐
𝟐
Now, we get:
𝒏 𝒏
𝟔 𝟏 𝟏 𝟒
𝒂𝒏 = ∑ = 𝟔∑( + − )=
𝒌(𝒌 + 𝟏)(𝟐𝒌 + 𝟏) 𝒌 𝒌 + 𝟏 𝟐𝒌 + 𝟏
𝒌=𝟏 𝒌=𝟏
𝒏 𝒏
𝟏 𝟏 𝟏 𝟏
= 𝟔∑( − ) − 𝟐𝟒 ∑ ( − )=
𝒌+𝟏 𝒌 𝟐𝒌 + 𝟏 𝟐𝒌
𝒌=𝟏 𝒌=𝟏
𝟏
= 𝟔( − 𝟏) − 𝟐𝟒(𝑯𝟐𝒏+𝟏 − 𝑯𝟐𝒏 − 𝟏).
𝒏+𝟏
𝐥𝐢𝐦 (𝑯𝒏 − 𝐥𝐨𝐠 𝒏) = 𝜸
𝒏→∞
𝟐𝒏 + 𝟏
𝑯𝟐𝒏+𝟏 − 𝑯𝟐𝒏 = (𝑯𝟐𝒏+𝟏 − 𝐥𝐨𝐠(𝟐𝒏 + 𝟏)) − (𝑯𝟐𝒏 − 𝐥𝐨𝐠 𝟐𝒏) + 𝐥𝐨𝐠
𝟐𝒏
𝐥𝐢𝐦 (𝑯𝟐𝒏+𝟏 − 𝑯𝟐𝒏 ) = 𝐥𝐨𝐠 𝟐 𝐚𝐧𝐝 𝐥𝐢𝐦 𝒂𝒏 = 𝟔(𝟑 − 𝟒 𝐥𝐨𝐠 𝟐)
𝒏→∞ 𝒏→∞
Therefore,
−𝟏 ∑𝒌
𝒊=𝟏 𝒊
𝟐
𝒏 𝒌
𝒆
𝛀 = 𝐥𝐢𝐦 ∑ (𝟏 + (∑ 𝒊𝟐 ) ) − 𝒏𝒆 = − ⋅ 𝟔(𝟑 − 𝟒 𝐥𝐨𝐠 𝟐) =
𝒏→∞ 𝟐
𝒌=𝟏 𝒊=𝟏
( )
www.ssmrmh.ro
= 𝟑𝒆(𝟒 𝐥𝐨𝐠 𝟐 − 𝟑)
SP.392 Let (𝒙𝒏 )𝒏≥𝟏 , (𝒚𝒏 )𝒏≥𝟏 be sequences of real numbers such that
𝒙𝟏 = 𝟏, 𝒚𝟏 = 𝒂𝟐 + 𝒂 + 𝟏,
∗
𝒂𝟐 + 𝒂𝒙𝒏 + 𝒙𝒏 𝒚𝒏 𝒂𝟐 + 𝒂𝒚𝒏 + 𝒙𝒏 𝒚𝒏
𝒂 > 1, ∀𝑛 ∈ ℕ 𝐚𝐧𝐝 𝒙𝒏+𝟏 = , 𝒚𝒏+𝟏 = . 𝐅𝐢𝐧𝐝:
𝒚𝒏 𝒙𝒏
𝐥𝐢𝐦 𝒙𝒏 , 𝐥𝐢𝐦 𝒚𝒏 .
𝒏→∞ 𝒏→∞

𝒂𝟐 + 𝒂𝒙𝒏 + 𝒙𝒏 𝒚𝒏 𝒂(𝒂 + 𝒙𝒏 ) + (𝒂 + 𝒙𝒏 )𝒚𝒏 − 𝒂𝒚𝒏 (𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 )
𝒙𝒏+𝟏 = = = −𝒂
𝒚𝒏 𝒚𝒏 𝒚𝒏
(𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 ) 𝟏 𝒚𝒏
⇒ 𝒂 + 𝒙𝒏+𝟏 = ⇒ =
𝒚𝒏 𝒂 + 𝒙𝒏+𝟏 (𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 )
𝟏 𝒙𝒏
𝐒𝐢𝐦𝐢𝐥𝐚𝐫𝐥𝐲, = 𝐭𝐡𝐞𝐧:
𝒂 + 𝒚𝒏+𝟏 (𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 )
𝟏 𝟏 𝒚𝒏 − 𝒙 𝒏 (𝒚𝒏 + 𝒂) − (𝒙𝒏 + 𝒂) 𝟏 𝟏
− = = = −
𝒂 + 𝒙𝒏+𝟏 𝒂 + 𝒚𝒏+𝟏 (𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 ) (𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 ) 𝒂 + 𝒙𝒏 𝒂 + 𝒚 𝒏
Namely,
𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
− = − =⋯= − =
𝒂 + 𝒙𝒏+𝟏 𝒂 + 𝒚𝒏+𝟏 𝒂 + 𝒙𝒏 𝒂 + 𝒚𝒏 𝒂 + 𝒙 𝟏 𝒂 + 𝒚𝟏
𝟏 𝟏 𝒂
= − 𝟐
=
𝒂 + 𝟏 ( 𝒂 + 𝟏) ( 𝒂 + 𝟏) 𝟐
𝟏 𝒂 𝟏 𝒂
= 𝟐
+ > ; (𝒙𝒏 > 0, 𝒚𝒏 > 0, ∀𝒏 ∈ ℕ∗ − 𝐢𝐧𝐝𝐮𝐜𝐭𝐢𝐨𝐧)
𝒂 + 𝒙𝒏+𝟏 (𝒂 + 𝟏) 𝒂 + 𝒚𝒏+𝟏 (𝒂 + 𝟏)𝟐
( 𝒂 + 𝟏) 𝟐 ( 𝒂 + 𝟏) 𝟐 𝟏
⇒ 𝒂 + 𝒙𝒏 < ⇒ 𝒙𝒏 < − 𝒂 = 𝟐 + < 3, ∀𝑎 > 1
𝒂 𝒂 𝒂
∗
⇒ 𝒙𝒏 < 3, ∀𝑛 ∈ ℕ ; ⇒ (𝒙𝒏 )𝒏≥𝟏 − is bounded ; (𝟏)
𝒂𝟐 + 𝒂𝒙𝒏 𝒂𝟐 + 𝒂𝒚𝒏
𝒙𝒏+𝟏 − 𝒙𝒏 = > 0 𝐚𝐧𝐝 𝒚𝒏+𝟏 − 𝒚𝒏 = > 0 ⇒ (𝒙𝒏 )𝒏≥𝟏 , (𝒚𝒏 )𝒏≥𝟏 ↗; (𝟐).
𝒚𝒏 𝒙𝒏
From (𝟏), (𝟐) we obtain that (𝒙𝒏 )𝒏≥𝟏 is convergent.
Let 𝒍 = 𝐥𝐢𝐦 𝒙𝒏 , 𝒍 ∈ (𝟏, 𝟑) and suppose that 𝒍′ = 𝐥𝐢𝐦 𝒚𝒏 ; 𝒍′ ∈ ℝ hence,
𝒏→∞ 𝒏→∞

www.ssmrmh.ro
𝟐 ′
𝒂 + 𝒂𝒍 + 𝒍𝒍
𝒍= ⇔ 𝒍𝒍′ = 𝒂𝟐 + 𝒂𝒍 + 𝒍𝒍′ ⇔ 𝒍 = −𝒂 𝐜𝐨𝐧𝐭𝐫𝐚𝐝𝐢𝐜𝐭𝐢𝐨𝐧! ⇒ 𝐥𝐢𝐦 𝒚𝒏 = ∞.
𝒍′ 𝒏→∞
𝟏 𝟏 𝒂 𝟏 𝒂
− = 𝟐
; (𝒚𝒏 → +∞) ⇒ =
𝒂 + 𝒙 𝒏 𝒂 + 𝒚𝒏 ( 𝒂 + 𝟏 ) 𝒂 + 𝒙𝒏 ( 𝒂 + 𝟏) 𝟐
(𝒂 + 𝟏)𝟐 ( 𝒂 + 𝟏) 𝟐 𝟐𝒂 + 𝟏 𝟏
⇔ 𝒂+𝒍= ⇔𝒍= −𝒂= ⇒ 𝐥𝐢𝐦 𝒙𝒏 = 𝟐 + .
𝒂 𝒂 𝒂 𝒏→∞ 𝒂
Solution 2 by Kamel Gandouli Rezgui-Tunisia
𝒂𝟐 + 𝒂𝒙𝒏 + 𝒚𝒏 𝒙𝒏 𝒂𝟐 + 𝒂𝒚𝒏 + 𝒚𝒏 𝒙𝒏
𝒙𝒏+𝟏 = , 𝒚𝒏+𝟏 =
𝒚𝒏 𝒙𝒏
𝟐𝒂𝟔𝟐 + 𝟐𝒂 + 𝟏
𝒙𝟐 = < 𝒚𝟐
𝒂𝟐 + 𝒂 + 𝟏
𝒂𝟐 + 𝒂𝒙𝒏 + 𝒚𝒏 𝒙𝒏 𝒂𝟐 + 𝒂𝒚𝒏 + 𝒚𝒏 𝒙𝒏
𝒙𝒏+𝟏 − 𝒚𝒏+𝟏 = − =
𝒚𝒏 𝒙𝒏
(𝒙𝒏 − 𝒚𝒏 )(𝒂𝟐 + 𝒂𝒚𝒏 + 𝒂𝒙𝒏 + 𝒙𝒏 𝒚𝒏 ) (𝒙𝒏 − 𝒚𝒏 )(𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 )
= =
𝒙 𝒏 𝒚𝒏 𝒙 𝒏 𝒚𝒏
By mathematical induction 𝒚𝒏 > 𝒙𝒏 , ∀𝒏 ∈ ℕ
Now, 𝒙𝟏 , 𝒙𝟐 > 0. Suppose 𝒙𝒏 > 0, 𝒚𝒏 > 0 then
𝒂𝟐 + 𝒂𝒙𝒏 + 𝒚𝒏 𝒙𝒏 𝒂𝟐 + 𝒂𝒙𝒏
𝒙𝒏+𝟏 = = + 𝒙𝒏 > 0
𝒚𝒏 𝒚𝒏
𝒂𝟐 + 𝒂𝒚𝒏 + 𝒚𝒏 𝒙𝒏 𝒂𝟐 + 𝒂𝒚𝒏
𝒚𝒏+𝟏 = = + 𝒚𝒏 > 0, ∀𝑛 ∈ ℕ
𝒙𝒏 𝒙𝒏
( 𝒂 + 𝟏) 𝟐 ( 𝒂 + 𝟏) 𝟐 𝟏
⇒ 𝒂 + 𝒙𝒏 < ⇒ 𝒙𝒏 < − 𝒂 = 𝟐 + < 3, ∀𝑎 > 1
𝒂 𝒂 𝒂
⇒ 𝒙𝒏 < 3, ∀𝑛 ∈ ℕ∗ ; ⇒ (𝒙𝒏 )𝒏≥𝟏 − is bounded ; (𝟏)
𝒚𝒏 𝒙𝒏
Let 𝒙 = 𝐥𝐢𝐦 𝒙𝒏 , 𝒙 ∈ (𝟏, 𝟑) and suppose that 𝒚 = 𝐥𝐢𝐦 𝒚𝒏 ; 𝒚 ∈ ℝ hence,
𝒏→∞ 𝒏→∞
𝒂𝟐 + 𝒂𝒙 + 𝒙𝒚
𝒙= ⇔ 𝒙𝒚 = 𝒂𝟐 + 𝒂𝒙 + 𝒙𝒚 ⇔ 𝒙 = −𝒂 𝐜𝐨𝐧𝐭𝐫𝐚𝐝𝐢𝐜𝐭𝐢𝐨𝐧! ⇒ 𝐥𝐢𝐦 𝒚𝒏 = ∞.
𝒚 𝒏→∞
𝟏 𝟏 𝒂 𝟏 𝒂
− = ; (𝒚 𝒏 → +∞) ⇒ =
𝒂 + 𝒙 𝒏 𝒂 + 𝒚𝒏 ( 𝒂 + 𝟏 ) 𝟐 𝒂 + 𝒙𝒏 ( 𝒂 + 𝟏) 𝟐
www.ssmrmh.ro
(𝒂 + 𝟏)𝟐 ( 𝒂 + 𝟏) 𝟐 𝟐𝒂 + 𝟏 𝟏
⇔ 𝒂+𝒙 = ⇔𝒙= −𝒂= ⇒ 𝐥𝐢𝐦 𝒙𝒏 = 𝟐 + .
Solution 3 by Ruxandra Daniela Tonilă-Romania
𝒂𝟐 + 𝒂𝒙𝒏 + 𝒙𝒏 𝒚𝒏 𝒂(𝒂 + 𝒙𝒏 ) + (𝒂 + 𝒙𝒏 )𝒚𝒏 − 𝒂𝒚𝒏 (𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 )
𝒙𝒏+𝟏 = = = −𝒂
𝒚𝒏 𝒚𝒏 𝒚𝒏
(𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 ) 𝟏 𝒚𝒏
⇒ 𝒂 + 𝒙𝒏+𝟏 = ⇒ =
𝒚𝒏 𝒂 + 𝒙𝒏+𝟏 (𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 )
𝟏 𝒙𝒏
𝐒𝐢𝐦𝐢𝐥𝐚𝐫𝐥𝐲, = 𝐭𝐡𝐞𝐧:
𝒂 + 𝒚𝒏+𝟏 (𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 )
𝟏 𝟏 𝒚𝒏 − 𝒙 𝒏 (𝒚𝒏 + 𝒂) − (𝒙𝒏 + 𝒂) 𝟏 𝟏
− = = = −
𝒂 + 𝒙𝒏+𝟏 𝒂 + 𝒚𝒏+𝟏 (𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 ) (𝒂 + 𝒙𝒏 )(𝒂 + 𝒚𝒏 ) 𝒂 + 𝒙𝒏 𝒂 + 𝒚 𝒏
Namely,
𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
− = − =⋯= − =
𝒂 + 𝒙𝒏+𝟏 𝒂 + 𝒚𝒏+𝟏 𝒂 + 𝒙𝒏 𝒂 + 𝒚𝒏 𝒂 + 𝒙 𝟏 𝒂 + 𝒚𝟏
𝟏 𝟏 𝒂
= − 𝟐
=
𝒂 + 𝟏 ( 𝒂 + 𝟏) ( 𝒂 + 𝟏) 𝟐
𝟏 𝒂 𝟏 𝒂
= 𝟐
+ > ; (𝒙𝒏 > 0, 𝒚𝒏 > 0, ∀𝒏 ∈ ℕ∗ − 𝐢𝐧𝐝𝐮𝐜𝐭𝐢𝐨𝐧)
𝒂 + 𝒙𝒏+𝟏 (𝒂 + 𝟏) 𝒂 + 𝒚𝒏+𝟏 (𝒂 + 𝟏)𝟐
( 𝒂 + 𝟏) 𝟐 ( 𝒂 + 𝟏) 𝟐 𝟏
⇒ 𝒂 + 𝒙𝒏 < ⇒ 𝒙𝒏 < − 𝒂 = 𝟐 + < 3, ∀𝑎 > 1
𝒂 𝒂 𝒂
∗
⇒ 𝒙𝒏 < 3, ∀𝑛 ∈ ℕ ; ⇒ (𝒙𝒏 )𝒏≥𝟏 − is bounded ; (𝟏)
𝒚𝒏 𝒙𝒏
Let 𝒙 = 𝐥𝐢𝐦 𝒙𝒏 , 𝒙 ∈ (𝟏, 𝟑) and suppose that 𝒚 = 𝐥𝐢𝐦 𝒚𝒏 ; 𝒚 ∈ ℝ hence,
𝒏→∞ 𝒏→∞
𝟐
𝒂 + 𝒂𝒙 + 𝒙𝒚
𝒙= ⇔ 𝒙𝒚 = 𝒂𝟐 + 𝒂𝒙 + 𝒙𝒚 ⇔ 𝒙 = −𝒂 𝐜𝐨𝐧𝐭𝐫𝐚𝐝𝐢𝐜𝐭𝐢𝐨𝐧! ⇒ 𝐥𝐢𝐦 𝒚𝒏 = ∞.
𝒚 𝒏→∞
𝟏 𝟏 𝒂 𝟏 𝒂
− = 𝟐
; (𝒚𝒏 → +∞) ⇒ =
𝒂 + 𝒙 𝒏 𝒂 + 𝒚𝒏 ( 𝒂 + 𝟏 ) 𝒂 + 𝒙𝒏 ( 𝒂 + 𝟏) 𝟐
(𝒂 + 𝟏)𝟐 ( 𝒂 + 𝟏) 𝟐 𝟐𝒂 + 𝟏 𝟏
⇔ 𝒂+𝒙 = ⇔𝒙= −𝒂= ⇒ 𝐥𝐢𝐦 𝒙𝒏 = 𝟐 + .

www.ssmrmh.ro
𝟏 𝟏 𝟏
+ +⋯+𝒏+𝒌⋅𝒏
SP.393 If 𝑺𝒏,𝒌 = 𝒆 𝒏 𝒏+𝟏 , 𝒏, 𝒌 ∈ ℕ, 𝒌 ≥ 𝟏 then find:
𝒏 𝒌
𝟏 𝑺𝒏,𝒌
𝛀 = 𝐥𝐢𝐦 (∑ (𝟏 + ) − 𝒏)
𝒏→∞ 𝒏
𝒌=𝟏

We want to prove the following inequality:
(∗) : 𝐥𝐨𝐠(𝟏 + 𝒙) < 𝑥 < − 𝐥𝐨𝐠(𝟏 − 𝒙) , ∀𝒙 ∈ (𝟎, 𝟏)
𝒙
Let 𝒇: [𝟎, 𝟏] → ℝ, 𝒇(𝒙) = 𝒙 − 𝐥𝐨𝐠(𝟏 + 𝒙) with 𝒇′ (𝒙) = 𝟏+𝒙 > 0, ∀𝑥 ∈ (𝟎, 𝟏) then,
𝒇 −increasing function on (𝟎, 𝟏). How, 𝒇(𝟎) = 𝟎 then, 𝒙 − 𝐥𝐨𝐠(𝟏 + 𝒙) > 0, ∀𝑥 ∈ (𝟎, 𝟏).
𝒙
Now, let be the function 𝒈: [𝟎, 𝟏) → ℝ, 𝒈(𝒙) = 𝒙 + 𝐥𝐨𝐠(𝟏 − 𝒙) with 𝒈′ (𝒙) = − 𝟏−𝒙 < 0,
∀𝒙 ∈ (𝟎, 𝟏) then, 𝒈 −decreasing on (𝟎, 𝟏) then, 𝒈(𝒙) < 𝑔(𝟎) = 𝟎, ∀𝒙 ∈ (𝟎, 𝟏).
𝟏 𝟏 𝟏
𝐋𝐞𝐭: 𝒙𝒏 = + +⋯+
𝒏 𝒏+𝟏 𝒏+𝒌⋅𝒏
Using the inequality (∗), it follows that:
𝟏 𝟏 𝟏
𝐥𝐨𝐠 (𝟏 + ) < < − 𝐥𝐨𝐠 (𝟏 − )
𝒏 𝒏 𝒏
𝟏 𝟏 𝟏
𝐥𝐨𝐠 (𝟏 + )< < − 𝐥𝐨𝐠 (𝟏 − )
𝒏+𝟏 𝒏+𝟏 𝒏+𝟏
⋮
𝟏 𝟏 𝟏
𝐥𝐨𝐠 (𝟏 + )< < − 𝐥𝐨𝐠 (𝟏 − )
𝒏+𝒌⋅𝒏 𝒏+𝒌⋅𝒏 𝒏 + 𝒌𝒏
Adding these relations, it follows that:
𝒌𝒏 𝒌𝒏
𝟏 𝟏
𝐥𝐨𝐠 (∏ (𝟏 + )) < 𝒙𝒏 < − 𝐥𝐨𝐠 (∏ (𝟏 − ))
𝒏+𝒊 𝒏+𝒊
𝒊=𝟎 𝒊=𝟎
𝒌𝒏 𝒌𝒏
𝟏 𝟏 𝟏 𝒏−𝟏
𝐁𝐮𝐭: ∏ (𝟏 + ) = 𝒌 + 𝟏 + 𝐚𝐧𝐝 ∏ (𝟏 − )= 𝐭𝐡𝐮𝐬,
𝒏+𝒊 𝒏 𝒏+𝒊 𝒏 + 𝒌𝒏
𝒊=𝟎 𝒊=𝟎
𝟏 𝒌+𝟏
𝐥𝐨𝐠 (𝒌 + 𝟏 + ) < 𝒙𝒏 < 𝐥𝐨𝐠 (𝒌 + 𝟏 + ) 𝐭𝐡𝐞𝐧,
𝒏 𝒏−𝟏

www.ssmrmh.ro
𝐥𝐢𝐦 𝒙
𝐥𝐢𝐦 𝒙𝒏 = 𝐥𝐨𝐠(𝒌 + 𝟏) ⇒ 𝐥𝐢𝐦 𝑺𝒏 = 𝒆𝒏→∞ 𝒏 = 𝒌 + 𝟏
𝒏→∞ 𝒏→∞
𝟏 𝟐 𝒏
𝟏 𝟐 𝟏 𝟑 𝟏 𝒏+𝟏
⇒ 𝛀 = 𝐥𝐢𝐦 ((𝟏 + ) + (𝟏 + ) + ⋯ + (𝟏 + ) − 𝒏)
𝒏→∞ 𝒏 𝒏 𝒏
𝒏 𝒌 𝒏 𝒌
𝟏 𝑺𝒏 𝟏 𝒌+𝟏
𝐋𝐞𝐭: 𝒚𝒏 = ∑ (𝟏 + ) − 𝒏; 𝒏, 𝒌 ≥ 𝟏 ⇒ 𝒚𝒏 = ∑ (𝟏 + ) − 𝒏; 𝒏 ≥ 𝟏.
𝒏 𝒏
𝒌=𝟏 𝒌=𝟏
We have:
𝒏
𝒏 𝟏 𝒏+𝟏
𝟏 𝒏+𝟏 (𝟏 + 𝒏) −𝟏
𝒚𝒏 ≤ 𝒏 (𝟏 + ) −𝒏 =
𝒏 𝟏
𝒏
𝒏
𝟏 𝒏+𝟏 𝟏
(𝟏 + 𝒏) −𝟏 (𝟏 + 𝒙)𝟏+𝒙 − 𝟏 𝑳′ 𝑯
𝐥𝐢𝐦 = 𝐥𝐢𝐦 = 𝟏; (𝟏)
𝒏→∞ 𝟏 𝒙→𝟎 𝒙
𝒏
𝟏
𝒕 𝟏
From Bernoulli inequality, (𝟏 + 𝒕)𝒎 ≤ 𝟏 + 𝒎 for 𝒕 = 𝒏 and 𝒎 = 𝒌 + 𝟏, we get:
𝟏
𝟏 𝒌+𝟏 𝟏
(𝟏 + ) ≤𝟏+
𝒏 𝒏(𝒌 + 𝟏)
Namely,
𝒌 𝟏 𝟏
𝟏 𝒌+𝟏 𝟏+ 𝟏+ 𝒏(𝒌 + 𝟏) + 𝒌 + 𝟏
(𝟏 + ) = 𝒏 ≥ 𝒏 =
𝒏 𝟏 𝟏 𝒏(𝒌 + 𝟏) + 𝟏
𝟏 𝒌+𝟏 𝟏+ (
(𝟏 + 𝒏) 𝒏 𝒌 + 𝟏)
It follows that:
𝒏 𝒏 𝒏
𝒏(𝒌 + 𝟏) + 𝒌 + 𝟏 𝒌 𝒌−𝟏
𝒚𝒏 ≥ ∑ ( − 𝟏) = ∑ ≥∑ =
𝒏(𝒌 + 𝟏) + 𝟏 𝒏(𝒌 + 𝟏) + 𝟏 𝒏𝒌
𝒌=𝟏 𝒌=𝟏 𝒌=𝟏
𝒏 𝒏
𝟏 𝟏 𝟏 𝟏 𝑯𝒏
∑( − ) = 𝟏− ∑ = 𝟏− → 𝟏; (𝟐)
𝒏 𝒏𝒌 𝒏 𝒌 𝒏
𝒌=𝟏 𝒌=𝟏
From (𝟏), (𝟐) it follows that: 𝐥𝐢𝐦 𝒂𝒏 = 𝟏

𝒏→∞
Therefore,

www.ssmrmh.ro
𝒏 𝒌
𝟏 𝑺𝒏,𝒌
𝛀 = 𝐥𝐢𝐦 (∑ (𝟏 + ) − 𝒏) = 𝟏
𝒏→∞ 𝒏
𝒌=𝟏

𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
𝑺𝒏,𝒌 = 𝒆(𝒏+𝒏+𝟏+⋯+𝟐𝒏)+(𝟐𝒏+𝟏+𝟐𝒏+𝟐+⋯+𝟑𝒏)+⋯+(𝒌𝒏+𝒌𝒏+𝟏+⋯.+𝒌𝒏+𝒏)
𝟏 𝒏
𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
𝐥𝐢𝐦 ( + + ⋯+ ) = 𝐥𝐢𝐦 ∑ = ∫ 𝒅𝒙 =
𝒏→∞ 𝒌𝒏 𝒌𝒏 + 𝟏 𝒌𝒏 + 𝒏 𝒏→∞ 𝒏 𝒊 𝟎 𝒌+𝒙
𝒊=𝟎 𝒌 +
𝒏
𝒌+𝟏
= 𝐥𝐨𝐠(𝒌 + 𝒙)|𝟏𝟎 = 𝐥𝐨𝐠 ( )
𝒌
Hence,
𝟑 𝒌−𝟏 𝟑𝟒 𝒌+𝟏
= 𝒆𝐥𝐨𝐠(𝟐⋅𝟐⋅𝟑⋅…⋅ )
𝐥𝐢𝐦 𝑺𝒏,𝒌 = 𝒆𝐥𝐨𝐠 𝟐+𝐥𝐨𝐠𝟐+⋯+𝐥𝐨𝐠 𝒌 𝒌 =𝒌+𝟏
𝒏→∞
𝟏 𝟐 𝒏
𝟏 𝟐 𝟏 𝟑 𝟏 𝒏+𝟏
⇒ 𝛀 = 𝐥𝐢𝐦 ((𝟏 + ) + (𝟏 + ) + ⋯ + (𝟏 + ) − 𝒏)
𝒏→∞ 𝒏 𝒏 𝒏
We have:
𝒏
𝒏 𝒌 𝒏 𝟏 𝒏+𝟏
𝟏 𝑺𝒏 𝟏 𝒏+𝟏 (𝟏 + 𝒏) −𝟏
∑ (𝟏 + ) − 𝒏 ≤ 𝒏 (𝟏 + ) −𝒏=
𝒏 𝒏 𝟏
𝒌=𝟏
𝒏
𝒏
𝟏 𝒏+𝟏 𝟏
(𝟏 + 𝒏) −𝟏 (𝟏 + 𝒙)𝟏+𝒙 − 𝟏 𝑳′ 𝑯
𝐥𝐢𝐦 = 𝐥𝐢𝐦 = 𝟏; (𝟏)
𝒏→∞ 𝟏 𝒙→𝟎 𝒙
𝒏
𝟏
𝒕 𝟏
From Bernoulli inequality, (𝟏 + 𝒕)𝒎 ≤ 𝟏 + 𝒎 for 𝒕 = 𝒏 and 𝒎 = 𝒌 + 𝟏, we get:
𝟏
𝟏 𝒌+𝟏 𝟏
(𝟏 + ) ≤𝟏+
𝒏 𝒏(𝒌 + 𝟏)
Namely,
𝒌 𝟏 𝟏
𝟏 𝒌+𝟏 𝟏+ 𝟏+ 𝒏(𝒌 + 𝟏) + 𝒌 + 𝟏
(𝟏 + ) = 𝒏 ≥ 𝒏 =
𝒏 𝟏 𝟏 𝒏(𝒌 + 𝟏) + 𝟏
𝟏 𝒌+𝟏 𝟏+ (
(𝟏 + 𝒏) 𝒏 𝒌 + 𝟏)

www.ssmrmh.ro
It follows that:
𝒏 𝒌 𝒏 𝒏 𝒏
𝟏 𝑺𝒏 𝒏(𝒌 + 𝟏) + 𝒌 + 𝟏 𝒌 𝒌−𝟏
∑ (𝟏 + ) − 𝒏 ≥ ∑ ( − 𝟏) = ∑ ≥∑ =
𝒏 𝒏(𝒌 + 𝟏) + 𝟏 𝒏(𝒌 + 𝟏) + 𝟏 𝒏𝒌
𝒌=𝟏 𝒌=𝟏 𝒌=𝟏 𝒌=𝟏
𝒏 𝒏
𝟏 𝟏 𝟏 𝟏 𝑯𝒏
∑( − ) = 𝟏− ∑ = 𝟏− → 𝟏; (𝟐)
𝒏 𝒏𝒌 𝒏 𝒌 𝒏
𝒌=𝟏 𝒌=𝟏
From (𝟏), (𝟐) it follows that: 𝐥𝐢𝐦 𝒂𝒏 = 𝟏

𝒏→∞
Therefore,
𝒏 𝒌
𝟏 𝑺𝒏,𝒌
𝛀 = 𝐥𝐢𝐦 (∑ (𝟏 + ) − 𝒏) = 𝟏
𝒏→∞ 𝒏
𝒌=𝟏
SP.394 If (𝑭𝒏 )𝒏≥𝟎 , 𝑭𝟎 = 𝑭𝟏 = 𝟏, 𝑭𝒏+𝟐 = 𝑭𝒏+𝟏 + 𝑭𝒏 , ∀𝒏 ∈ ℕ and 𝒂, 𝒃, 𝒄 ∈ ℝ∗+

with 𝒂 + 𝒃 + 𝒄 ≤ 𝟐𝟒, then prove:
𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐
+ + ≥ 𝟏, ∀𝒏 ∈ ℕ
√𝑭𝟐𝒏 + 𝒂𝑭𝒏+𝟏 𝑭𝒏+𝟐
√𝑭𝟐𝒏+𝟏 + 𝒃𝑭𝒏+𝟐 𝑭𝒏 √𝑭𝟐𝒏+𝟐 + 𝒄𝑭𝒏 𝑭𝒏+𝟏
Proposed by D.M. Bătineţu-Giurgiu, Neculai Stanciu-Romania

𝐋𝐞𝐭: 𝑼𝒏 = + +
√𝑭𝟐𝒏 + 𝒂𝑭𝒏+𝟏 𝑭𝒏+𝟐 √𝑭𝟐𝒏+𝟏 + 𝒃𝑭𝒏+𝟐 𝑭𝒏 √𝑭𝟐𝒏+𝟐 + 𝒄𝑭𝒏 𝑭𝒏+𝟏
𝑭𝒏 𝑭𝟐𝒏
𝑼𝒏 = ∑ =∑ =
𝒄𝒚𝒄
√𝑭𝟐𝒏 + 𝒂𝑭𝒏+𝟏 𝑭𝒏+𝟐 𝑭 ⋅ √𝑭𝟑𝒏 + 𝒂𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐
𝒄𝒚𝒄 √ 𝒏
𝑭𝟐𝒏
=∑ , ∀𝒏 ∈ ℕ, 𝐰𝐡𝐞𝐫𝐞 𝒑𝒏 = 𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐 , ∀𝒏 ∈ ℕ; (𝟏)
𝒄𝒚𝒄 √𝑭𝒏 ⋅ √𝑭𝟑𝒏 + 𝒂𝒑𝒏
We denote: 𝒗𝒏 = √𝑭𝒏 ⋅ √𝑭𝟑𝒏 + 𝒂𝒑𝒏 and 𝑽𝒏 = 𝒗𝒏 + 𝒗𝒏+𝟏 + 𝒗𝒏+𝟐 =
= √𝑭𝒏 ⋅ √𝑭𝟑𝒏 + 𝒂𝒑𝒏 + √𝑭𝒏+𝟏 ⋅ √𝑭𝟑𝒏+𝟏 + 𝒂𝒑𝒏+𝟏 + √𝑭𝒏+𝟐 ⋅ √𝑭𝟑𝒏+𝟐 + 𝒂𝒑𝒏+𝟐 , ∀𝒏 ∈ ℕ.

www.ssmrmh.ro
𝟐 𝑪𝑩𝑺
We have: 𝑽𝟐𝒏 = (∑ √𝑭𝒏 ⋅ √𝑭𝟑𝒏 + 𝒂𝒑𝒏 ) ≤ 𝒔𝒏 (𝑭𝟑𝒏 + 𝑭𝟑𝒏+𝟏 + 𝑭𝟑𝒏+𝟐 + (𝒂 + 𝒃 + 𝒄)𝒑𝒏 ), ∀𝒏 ∈
ℕ, where we denote 𝒔𝒏 = 𝑭𝒏 + 𝑭𝒏+𝟏 + 𝑭𝒏+𝟐 , ∀𝒏 ∈ ℕ.

Since (𝒙 + 𝒚 + 𝒛)𝟑 ≥ 𝒙𝟑 + 𝒚𝟑 + 𝒛𝟑 + 𝟐𝟒𝒙𝒚𝒛, ∀𝒙, 𝒚, 𝒛 ∈ ℝ∗+ and
𝑽𝟐𝒏 ≤ 𝒔𝒏 (𝑭𝟑𝒏 + 𝑭𝟑𝒏+𝟏 + 𝑭𝟑𝒏+𝟐 + (𝒂 + 𝒃 + 𝒄)𝒑𝒏 )
≤ 𝒔𝒏 (𝑭𝟑𝒏 + 𝑭𝟑𝒏+𝟏 + 𝑭𝟑𝒏+𝟐 + 𝟐𝟒𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐 ),
then 𝑽𝟐𝒏 ≤ 𝒔𝒏 ⋅ 𝒔𝟑𝒏 = 𝒔𝟒𝒏 , ∀𝒏 ∈ ℕ ⇔ 𝑽𝒏 ≤ 𝒔𝟐𝒏 , ∀𝒏 ∈ ℕ; (𝟐)
Applying Bergstrom’s inequality, and from (𝟏), (𝟐) it follows that:
𝒔𝟐𝒏 𝒔𝟐𝒏
𝑼𝒏 ≥ ≥ = 𝟏, ∀𝒏 ∈ ℕ.
𝑽𝒏 𝒔𝟐𝒏
+ + =
√𝑭𝟐𝒏 + 𝒂𝑭𝒏+𝟏 𝑭𝒏+𝟐 √𝑭𝟐𝒏+𝟏 √𝑭𝟐𝒏+𝟐
+ 𝒃𝑭𝒏 𝑭𝒏+𝟐 + 𝒄𝑭𝒏 𝑭𝒏+𝟏
𝟑 𝟑 𝟑
𝑭𝟐𝒏 𝑭𝟐𝒏+𝟏 𝑭𝟐𝒏+𝟐 𝑹𝒂𝒅𝒐𝒏
= + + ≥
√𝑭𝟑𝒏 + 𝒂𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐 √𝑭𝟑𝒏+𝟏 √𝑭𝟑𝒏+𝟐
+ 𝒃𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐 + 𝒄𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐
√(𝑭𝒏 + 𝑭𝒏+𝟏 + 𝑭𝒏+𝟐 )𝟑
= =
√𝑭𝟑𝒏 + 𝑭𝟑𝒏+𝟏 + 𝑭𝟑𝒏+𝟐 + (𝒂 + 𝒃 + 𝒄)𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐
√𝑭𝟑𝒏 + 𝑭𝟑𝒏+𝟏 + 𝑭𝟑𝒏+𝟐 + 𝟑(𝑭𝒏 + 𝑭𝒏+𝟏 )(𝑭𝒏+𝟏 + 𝑭𝒏+𝟐 )(𝑭𝒏+𝟐 + 𝑭𝒏 ) 𝑨𝑴−𝑮𝑴

= ≥
𝟑 𝟑 𝟑
√𝑭𝒏 + 𝑭𝒏+𝟏 + 𝑭𝒏+𝟐 + (𝒂 + 𝒃 + 𝒄)𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐
√𝑭𝟑𝒏 + 𝑭𝟑𝒏+𝟏 + 𝑭𝟑𝒏+𝟐 + 𝟐𝟒𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐

≥ ≥𝟏
√𝑭𝟑𝒏 + 𝑭𝟑𝒏+𝟏 + 𝑭𝟑𝒏+𝟐 + (𝒂 + 𝒃 + 𝒄)𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐
Since 𝒂 + 𝒃 + 𝒄 ≤ 𝟐𝟒.
Solution 3 by Amrit Awasthi-India
𝑭𝒏 𝑭𝒏 √𝑭𝒏
𝛀=∑ =∑ =
𝒄𝒚𝒄
√𝑭𝟐𝒏 + 𝒂𝑭𝒏+𝟏 𝑭𝒏+𝟐 𝒄𝒚𝒄
√𝑭𝟑𝒏 + 𝒂𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐

www.ssmrmh.ro
𝑭𝟐𝒏 𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎 (𝑭𝒏 + 𝑭𝒏+𝟏 + 𝑭𝒏+𝟐 )𝟐
=∑ ≥ =
𝑭 ⋅ √𝑭𝟑𝒏 + 𝒂𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐
𝒄𝒚𝒄 √ 𝒏
∑ √𝑭𝒏 ⋅ √𝑭𝟑𝒏 + 𝒂𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐
𝟒𝑭𝟐𝒏+𝟐
= ; (∗)
√𝑭𝒏 ⋅ √𝑭𝟑𝒏 + 𝒂𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐
Now apply Cauchy-Schwarz in the term in denominator, we get
𝟐
(∑ √𝑭𝒏 ⋅ √𝑭𝟑𝒏 + 𝒂𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐 ) ≤ (∑ 𝑭𝒏 ) (∑(𝑭𝟑𝒏 + 𝒂𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐 ))
𝟐
(∑ √𝑭𝒏 ⋅ √𝑭𝟑𝒏 + 𝒂𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐 )
≤ (∑ 𝑭𝒏 )(𝑭𝟑𝒏 + 𝑭𝟑𝒏+𝟏 + 𝑭𝟑𝒏+𝟐 + (𝒂 + 𝒃 + 𝒄)𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐
≤ (∑ 𝑭𝒏 ) (𝑭𝟑𝒏 + 𝑭𝟑𝒏+𝟏 + 𝑭𝟑𝒏+𝟐 + 𝟐𝟒𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐 ); (𝒂 + 𝒃 + 𝒄 = 𝟐𝟒)

𝟑 𝟒
≤ (∑ 𝑭𝒏 ) (∑ 𝑭𝒏 ) = (∑ 𝑭𝒏 )
∵ (𝒙 + 𝒚 + 𝒛)𝟑 ≥ 𝒙𝟑 + 𝒚𝟑 + 𝒛𝟑 + 𝟐𝟒𝒙𝒚𝒛
𝟏 𝟏 𝟏
≥ =
∑ √𝑭𝒏 ⋅ √𝑭𝟑𝒏 + 𝒂𝑭𝒏 𝑭𝒏+𝟏 𝑭𝒏+𝟐 (∑ 𝑭𝒏 ) 𝟐
𝟒𝑭𝟐𝒏+𝟐
𝟒𝑭𝟐
Substitute in (∗) we get 𝛀 ≥ 𝟒𝑭𝒏+𝟐
𝟐 = 𝟏.Therefore,
𝒏+𝟐
𝑭𝒏
∑ ≥𝟏
𝒄𝒚𝒄
√𝑭𝟐𝒏 + 𝒂𝑭𝒏+𝟏 𝑭𝒏+𝟐
SP.395 𝑨 ∈ 𝑴𝟐 (ℚ) such that 𝐝𝐞𝐭(𝑨𝟐 − 𝟐𝑰𝟐 ) = 𝟎. Find:

𝛀 = 𝐝𝐞𝐭(𝑨𝟐 − 𝟑𝑨 + 𝟑𝑰𝟐 )
Let 𝒇(𝒙) = 𝐝𝐞𝐭(𝑨 + 𝒙𝑰𝟐 ) = 𝐝𝐞𝐭 𝑨 + 𝒂𝟏 𝒙 + 𝒙𝟐 , 𝒂𝟏 ∈ ℚ; (𝟏)
𝐝𝐞𝐭(𝑨𝟐 − 𝟐𝑰𝟐 ) = 𝟎 ⇔ 𝐝𝐞𝐭(𝑨 + √𝟐𝑰𝟐 ) ⋅ 𝐝𝐞𝐭(𝑨 − √𝟐𝑰𝟐 ) = 𝟎 ⇔
𝐝𝐞𝐭(𝑨 + √𝟐𝑰𝟐 ) = 𝟎 or𝐝𝐞𝐭(𝑨 − √𝟐𝑰𝟐 ) = 𝟎 ⇒ 𝒇(√𝟐) = 𝟎 or 𝒇(−√𝟐) = 𝟎; (𝟐)

From (𝟏), (𝟐) it follows that:
𝐝𝐞𝐭 𝑨 ± 𝒂𝟏 √𝟐 + 𝟐 = 𝟎 ⇒ 𝒂𝟏 = 𝟎 ⇒ 𝐝𝐞𝐭 𝑨 = −𝟐 ⇒ 𝒇(𝒙) = 𝐝𝐞𝐭(𝑨 + 𝒙𝑰𝟐 ) = 𝒙𝟐 − 𝟐; (𝟑)
www.ssmrmh.ro
𝟐
𝛀 = 𝐝𝐞𝐭 𝑨 − 𝟑𝑨 + 𝟑𝑰𝟐 ) = 𝐝𝐞𝐭(𝑨 + 𝒙𝟏 𝑰𝟐 )(𝑨 + 𝒙𝟐 𝑰𝟐 ) =
(
= 𝐝𝐞𝐭(𝑨𝟐 + (𝒙𝟏 + 𝒙𝟐 )𝑨 + 𝒙𝟏 𝒙𝟐 𝑰𝟐 ) = 𝐝𝐞𝐭(𝑨𝟐 − 𝟑𝑨 + 𝟑𝑰𝟐 ) ; (𝟒),where𝒙𝟏,𝟐 − are roots of
the equation 𝒙𝟐 + 𝟑𝒙 + 𝟑 = 𝟎 and 𝒙𝟏 + 𝒙𝟐 = −𝟑, 𝒙𝟏 𝒙𝟐 = 𝟑.
From (𝟑), (𝟒) we have:
𝛀 = 𝐝𝐞𝐭(𝑨𝟐 − 𝟑𝑨 + 𝟑𝑰𝟐 ) = 𝒇(𝒙𝟏 ) ⋅ 𝒇(𝒙𝟐 ) = (𝒙𝟐𝟏 − 𝟐)(𝒙𝟐𝟐 − 𝟐) =
= (𝒙𝟏 𝒙𝟐 )𝟐 − 𝟐(𝒙𝟐𝟏 + 𝒙𝟐𝟐 ) + 𝟒 = (𝒙𝟏 𝒙𝟐 )𝟐 − 𝟐[(𝒙𝟏 + 𝒙𝟐 )𝟐 − 𝟐𝒙𝟏 𝒙𝟐 ] + 𝟒 =
= 𝟗 − 𝟐(𝟗 − 𝟔) + 𝟒 = 𝟕.
𝒂 𝒃
Let 𝑨 = ( ) ; 𝒂, 𝒃, 𝒄, 𝒅 ∈ ℚ, 𝐝𝐞𝐭(𝑨𝟐 − 𝟐𝑰𝟐 ) = 𝟎
𝒄 𝒅
⇒ 𝐝𝐞𝐭(𝑨 − √𝟐𝑰𝟐 )(𝑨 + √𝟐𝑰𝟐 ) = 𝟎 ⇒ 𝐝𝐞𝐭(𝑨 − √𝟐𝑰𝟐 ) ⋅ 𝐝𝐞𝐭(𝑨 + √𝟐𝑰𝟐 ) = 𝟎
⇒ 𝐝𝐞𝐭(𝑨 − √𝟐𝑰𝟐 ) = 𝟎 𝐨𝐫 𝐝𝐞𝐭(𝑨 + √𝟐𝑰𝟐 ) = 𝟎

𝒂 − √𝟐 𝒃
Assume that 𝐝𝐞𝐭(𝑨 − √𝟐𝑰𝟐 ) = 𝟎 ⇒ | |=𝟎
𝒄 𝒅 − √𝟐
⇒ (𝒂 − √𝟐)(𝒅 − √𝟐) − 𝒃𝒄 = 𝟎 ⇒ (𝒂𝒅 − 𝒃𝒄 + 𝟐) − √𝟐(𝒂 + 𝒅) = 𝟎
As 𝒂, 𝒃, 𝒄, 𝒅 ∈ ℚ, 𝒂𝒅 − 𝒃𝒄 + 𝟐 = 𝟎 and 𝒂 + 𝒅 = 𝟎
∵ 𝒂𝒅 − 𝒃𝒄 = −𝟐
By the Cauchy-Hamilton Theorem, we have
𝑨𝟐 − (𝒂 + 𝒅)𝑨 + (𝒂𝒅 − 𝒃𝒄)𝑰𝟐 = 𝑶𝟐 ⇒ 𝑨𝟐 − 𝟐𝑰𝟐 = 𝑶𝟐 𝐨𝐫 𝑨𝟐 = 𝟐𝑰𝟐
𝟓 − 𝟑𝒂 −𝟑𝒃
∵ 𝑨𝟐 − 𝟑𝑨 + 𝟑𝑰𝟐 = 𝟐𝑰𝟐 − 𝟑𝑨 + 𝟑𝑰𝟐 = ( )
−𝟑𝒄 𝟓 − 𝟑𝒅
𝐝𝐞𝐭(𝑨𝟐 − 𝟑𝑨 + 𝟑𝑰𝟐 ) = (𝟓 − 𝟑𝒂)(𝟓 − 𝟑𝒅) − 𝟗𝒃𝒄 =
= 𝟐𝟓 − 𝟏𝟓(𝒂 + 𝒅) + 𝟗(𝒂𝒅 − 𝒃𝒄) = 𝟕
Solution 3 by Ruxandra Daniela Tonilă-Romania
𝐝𝐞𝐭(𝑨𝟐 − 𝟐𝑰𝟐 ) = 𝐝𝐞𝐭(𝑨 − √𝟐𝑰𝟐 )(𝑨 + √𝟐𝑰𝟐 ) = 𝐝𝐞𝐭(𝑨 − √𝟐𝑰𝟐 ) ⋅ 𝐝𝐞𝐭(𝑨 + √𝟐𝑰𝟐 ) = 𝟎
⇒ 𝐝𝐞𝐭(𝑨 − √𝟐𝑰𝟐 ) = 𝟎 𝐨𝐫 𝐝𝐞𝐭(𝑨 + √𝟐𝑰𝟐 ) = 𝟎
Let 𝝀𝟏 , 𝝀𝟐 be the eigenvalues of 𝑨 ⇒ 𝝀𝟏 = √𝟐 or 𝝀𝟐 = −√𝟐

www.ssmrmh.ro
𝑻𝒓(𝑨) = 𝝀𝟏 + 𝝀𝟐 ∈ ℚ
𝑨 ∈ 𝑴𝟐 (ℚ) ⇔ { 𝐝𝐞𝐭(𝑨) = 𝝀𝟏 ⋅ 𝝀𝟐 ∈ ℚ ⇔ 𝝀𝟐 = −𝝀𝟏 .
𝝀𝟏 ∈ ℝ\ℚ
So, 𝝀𝟏 = √𝟐 and 𝝀𝟐 = −√𝟐 ⇒ 𝑻𝒓(𝑨) = 𝟎, 𝐝𝐞𝐭(𝑨) = −𝟐

𝑷𝑨 (𝒙) = 𝒙𝟐 − 𝑻𝒓(𝑨) ⋅ 𝒙 + 𝐝𝐞𝐭(𝑨) = 𝒙𝟐 − 𝟐
𝟑±𝒊√𝟑
Let be the equation 𝒙𝟐 − 𝟑𝒙 + 𝟑 = 𝟎, 𝚫 = −𝟑, 𝒙𝟏,𝟐 = 𝟐
𝐝𝐞𝐭(𝑨𝟐 − 𝟑𝑨 + 𝟑𝑰𝟐 ) = 𝐝𝐞𝐭(𝑨 − 𝒙𝟏 𝑰𝟐 ) ⋅ 𝐝𝐞𝐭(𝑨 − 𝒙𝟐 𝑰𝟐 ) ⇔

𝟐 𝟐
𝟐
𝟑 + 𝒊√𝟑 𝟑 − 𝒊√𝟑
𝐝𝐞𝐭(𝑨 − 𝟑𝑨 + 𝟑𝑰𝟐 ) = 𝑷𝑨 (𝒙𝟏 ) ⋅ 𝑷𝑨 (𝒙𝟐 ) = (( ) − 𝟐) (( ) − 𝟐) =
𝟐 𝟐
𝟔 + 𝟔𝒊√𝟐 𝟔 − 𝟔𝒊√𝟐 −𝟐 + 𝟔𝒊√𝟑 −𝟐 − 𝟔𝒊√𝟑

=( − 𝟐) ( − 𝟐) = ⋅ =𝟕
𝟒 𝟒 𝟒 𝟒
SP.396 Find all continuous functions 𝒇: ℝ → ℝ such that

𝒇(𝒙) − 𝒇(𝒂𝒙) + 𝒇(𝒂𝟐 𝒙) − 𝒇(𝒂𝟑 𝒙) + 𝒇(𝒂𝟒 𝒙) = 𝒙, ∀𝒙 ∈ ℝ, 𝒂 ∈ (𝟎, 𝟏)
𝒙 → 𝒂𝒙 ⇒ 𝒇(𝒂) − 𝒇(𝒂𝟐 𝒙) + 𝒇(𝒂𝟑 𝒙) − 𝒇(𝒂𝟒 𝒙) + 𝒇(𝒂𝟓 𝒙) = 𝒂𝒙
⇒ 𝒇(𝒙) + 𝒇(𝒂𝟓 𝒙) = (𝟏 + 𝒂)𝒙
(−)
𝒙 → 𝒂𝟓 𝒙 ⇒ 𝒇(𝒂𝟓 𝒙) + 𝒇(𝒂𝟏𝟎 𝒙) = (𝟏 + 𝒂)𝒂𝟓 𝒙 ⇒ 𝒇(𝒙) − 𝒇(𝒂𝟏𝟎 𝒙) = (𝟏 + 𝒂)(𝟏 − 𝒂𝟓 )𝒙
𝒙 → 𝒂𝟏𝟎 𝒙 ⇒ 𝒇(𝒂𝟏𝟎 𝒙) − 𝒇(𝒂𝟐𝟎 𝒙) = (𝟏 + 𝒂)(𝟏 − 𝒂𝟓 )𝒂𝟏𝟎 𝒙
⋮
(+)
𝒙 → 𝒂𝟏𝟎 𝒙 ⇒ 𝒇(𝒂𝟏𝟎(𝒏−𝟏) 𝒙) − 𝒇(𝒂𝟏𝟎𝒏 𝒙) = (𝟏 + 𝒂)(𝟏 − 𝒂𝟓 )𝒂𝟏𝟎(𝒏−𝟏) 𝒙 ⇒
𝒇(𝒙) − 𝒇(𝒂𝟏𝟎𝒏 𝒙) = (𝟏 + 𝒂)(𝟏 − 𝒂𝟓 )𝒙(𝟏 + 𝒂𝟏𝟎 + ⋯ + 𝒂𝟏𝟎(𝒏−𝟏) )
𝒂𝟏𝟎𝒏 − 𝟏
𝐥𝐢𝐦 (𝒇(𝒙) − 𝒇(𝒂𝟏𝟎𝒏 𝒙)) = 𝐥𝐢𝐦 (𝟏 + 𝒂)(𝟏 − 𝒂𝟓 )𝒙 ⋅ ⇒
𝒏→∞ 𝒏→∞ 𝒂𝟏𝟎 − 𝟏
(𝟏 + 𝒂)(𝟏 − 𝒂𝟓 )𝒙
𝒇(𝒙) − 𝒇(𝐥𝐢𝐦 𝒂𝟏𝟎𝒏 𝒙) =
𝒏→∞ 𝟏 − 𝒂𝟏𝟎
(𝟏 + 𝒂)𝒙
𝒇(𝒙) − 𝒇(𝟎) = 𝒙
{ 𝟏 + 𝒂𝟓 ⇒ 𝒇(𝒙) =
𝟏 − 𝒂 + 𝒂 − 𝒂𝟑 + 𝒂𝟒
𝟐
𝒇 ( 𝟎) = 𝟎
www.ssmrmh.ro
𝑳𝒆𝒕 𝒇(𝒙) − 𝒇(𝒂𝒙) + 𝒇(𝒂𝟐 𝒙) − 𝒇(𝒂𝟑 𝒙) + 𝒇(𝒂𝟒 𝒙) = 𝒙 (𝟏)
𝒙 = 𝟎 → 𝒇(𝟎) = 𝟎 𝒂𝒏𝒅 𝒙 ~ 𝒂𝒙 →
𝒇(𝒂𝒙) − 𝒇(𝒂𝟐 𝒙) + 𝒇(𝒂𝟑 𝒙) − 𝒇(𝒂𝟒 𝒙) + 𝒇(𝒂𝟓 𝒙) = 𝒂𝒙 (𝟐)
(𝟏) + (𝟐) → 𝒇(𝒙) + 𝒇(𝒂𝟓 𝒙) = (𝒂 + 𝟏)𝒙 (𝟑)
𝒙 ~ 𝒂𝟓 𝒙 𝒊𝒏 (𝟑) → 𝒇(𝒂𝟓 𝒙) + 𝒇(𝒂𝟏𝟎 𝒙) = (𝒂 + 𝟏)𝒂𝟓 𝒙 (𝟒)

(𝟒) − (𝟑) → 𝒇(𝒂𝟏𝟎 𝒙) − 𝒇(𝒙) = (𝒂 + 𝟏)(𝒂𝟓 − 𝟏)𝒙 (𝟓)
𝒙 ~ 𝒂𝟏𝟎𝒌 𝒙, 𝒌 ∈ 𝑵 𝒊𝒏 (𝟓) → 𝒇(𝒂𝟏𝟎(𝒌+𝟏) 𝒙) − 𝒇(𝒂𝟏𝟎𝒌 𝒙) = (𝒂 + 𝟏)(𝒂𝟓 − 𝟏)𝒂𝟏𝟎𝒌 𝒙

𝒏−𝟏 𝒏−𝟏
𝟏𝟎(𝒌+𝟏) 𝟏𝟎𝒌
→ ∑[𝒇(𝒂 𝒙) − 𝒇(𝒂 𝒙)] = (𝒂 + 𝟏)(𝒂 − 𝟏)𝒙 ∑ 𝒂𝟏𝟎𝒌 , 𝒏 ∈ 𝑵∗
𝟓
𝒌=𝟎 𝒌=𝟎
(𝒂 + 𝟏)(𝒂𝟓 − 𝟏)(𝟏 − 𝒂𝟏𝟎𝒏 )

↔ 𝒇(𝒂𝟏𝟎𝒏 𝒙) − 𝒇(𝒙) = 𝒙
𝟏 − 𝒂𝟏𝟎
𝒏→∞,𝒂∈(𝟎,𝟏)
𝒂+𝟏 𝒙
⏞
→ 𝒇(𝟎) − 𝒇(𝒙) = − 𝒙 = −
𝒂𝟓 + 𝟏 𝒂𝟒 − 𝒂𝟑 + 𝒂𝟐 − 𝒂 + 𝟏
𝒙
𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆, 𝒇(𝒙) = 𝟒 , ∀𝒙 ∈ 𝑹
𝒂 − 𝒂𝟑 + 𝒂𝟐 − 𝒂 + 𝟏
𝒇(𝒙) − 𝒇(𝒂𝒙) + 𝒇(𝒂𝟐 𝒙) − 𝒇(𝒂𝟑 𝒙) + 𝒇(𝒂𝟒 𝒙) = 𝒙; (𝟏)
Put 𝒙 = 𝟎 to obtain 𝒇(𝟎) = 𝟎. In (1) replace 𝒙 by 𝒂𝒙 so that
𝒇(𝒂𝒙) − 𝒇(𝒂𝟐 𝒙) + 𝒇(𝒂𝟑 𝒙) − 𝒇(𝒂𝟒 𝒙) + 𝒇(𝒂𝟓 𝒙) = 𝒂𝒙; (𝟐)
Adding (1) and (2) we get:
𝒇(𝒙) + 𝒇(𝒂𝟓 𝒙) = (𝒂 + 𝟏)𝒙
𝒇(𝒂𝟓 𝒙) + 𝒇(𝒂𝟏𝟎 𝒙) = (𝒂 + 𝟏)𝒂𝟓 𝒙
𝒇(𝒂𝟏𝟎 𝒙) + 𝒇(𝒂𝟏𝟓 𝒙) = (𝒂 + 𝟏)𝒙
……………………………………
𝒇(𝒂𝟓𝒏−𝟓 𝒙) + 𝒇(𝒂𝟓𝒏 𝒙) = (𝒂 + 𝟏)𝒂𝟓𝒏−𝟓 𝒙
Multiply the 2nd equation by (−𝟏), 4th equation by (−𝟏)𝟑 , 6th equation by (−𝟏)𝟓 and so
adding, we get

www.ssmrmh.ro
𝒏−𝟏
( ) (
𝒇 𝒙 + −𝟏 ) 𝒇(𝒂 𝒙) = 𝒂 + 𝟏)[𝟏 − 𝒂𝟓 + 𝒂𝟏𝟎 + ⋯ + (−𝟏)𝒏−𝟏 𝒂𝟓𝒏−𝟓 ]𝒙; (𝟑)
𝟓𝒏 (
𝒏→∞
As 𝟎 < 𝑎 < 1, 𝒂𝟓𝒏 → 𝟎 as 𝒏 → ∞ . Since 𝒇 −is continuous, 𝒇(𝒂𝟓𝒏 𝒙) → 𝒇(𝟎) = 𝟎.
Taking limit as 𝒏 → ∞ in (3) we get:
𝒙 ( 𝒂 + 𝟏)
𝒇(𝒙) = ; ∀𝒙 ∈ ℝ.
𝟏 + 𝒂𝟓
SP.397 Find:
𝒑 𝒎 𝒏
𝟏 𝒌𝟐
𝛀 = 𝐥𝐢𝐦 𝒂 ⋅ ∑ ∑ ∑ 𝟐
𝒑→∞ 𝒑 𝟐𝒌 − 𝟐𝒏𝒌 + 𝒏𝟐
𝒎=𝟏 𝒏=𝟏 𝒌=𝟏

𝒌𝟐
Let 𝒂𝒌 = 𝟐𝒌𝟐 −𝟐𝒏𝒌+𝒏𝟐 . Therefore,
𝒌𝟐 (𝒏 − 𝒌 )𝟐
𝒂𝒌 + 𝒂𝒏−𝒌 = + =
𝟐𝒌𝟐 − 𝟐𝒏𝒌 + 𝒏𝟐 𝟐(𝒏 − 𝒌)𝟐 − 𝟐𝒏(𝒏 − 𝒌) + 𝒏𝟐
𝒌𝟐 (𝒏 − 𝒌 )𝟐
= 𝟐 + =
𝟐𝒌 − 𝟐𝒏𝒌 + 𝒏𝟐 (𝒏 − 𝒌)[𝟐(𝒏 − 𝒌) − 𝟐𝒏] + 𝒏𝟐
𝒌𝟐 (𝒏 − 𝒌 ) 𝟐 𝒌𝟐 (𝒏 − 𝒌 )𝟐
= + = + =
𝟐𝒌𝟐 − 𝟐𝒏𝒌 + 𝒏𝟐 −𝟐𝒌(𝒏 − 𝒌) + 𝒏𝟐 𝟐𝒌𝟐 − 𝟐𝒏𝒌 + 𝒏𝟐 𝒏𝟐 − 𝟐𝒌𝒏 + 𝟐𝒌𝟐
𝒌 𝟐 + (𝒏 − 𝒌 ) 𝟐
= 𝟐 =𝟏
𝒏 − 𝟐𝒌𝒏 + 𝟐𝒌𝟐
Hence,
𝒏−𝟏 𝒏−𝟏
𝟐 ∑ 𝒂𝒌 = ∑ (𝒂𝒌 + 𝒂𝒏−𝒌 ) = 𝒏 − 𝟏
𝒌=𝟏 𝒌=𝟏
and then
𝒏
𝒌𝟐 𝒏+𝟏
∑ 𝟐 =
𝟐𝒌 − 𝟐𝒏𝒌 + 𝒏𝟐 𝟐
𝒌=𝟏
𝒑 𝒎 𝒏 𝒑 𝒎 𝒑
𝒌𝟐 𝒏+𝟏 𝒎(𝒎 + 𝟏) 𝒎
∑ ∑∑ 𝟐 = ∑ ∑ = ∑ ( + )=
𝟐𝒌 − 𝟐𝒏𝒌 + 𝒏𝟐 𝟐 𝟒 𝟐
𝒎=𝟏 𝒏=𝟏 𝒌=𝟏 𝒎=𝟏 𝒏=𝟏 𝒎=𝟏

www.ssmrmh.ro
𝒑 𝒑
𝟏 𝟏 𝟏 𝒑(𝒑 + 𝟏)(𝟐𝒑 + 𝟏) 𝟏 𝒑(𝒑 + 𝟏)
= ∑ 𝒎𝟐 + ∑ 𝒎 = ⋅ + ⋅ =
𝟒 𝟐 𝟒 𝟔 𝟐 𝟐
𝒎=𝟏 𝒎=𝟏
𝒑(𝒑 + 𝟏) 𝟐𝒑 + 𝟏 𝒑(𝒑 + 𝟏)(𝟐𝒑 + 𝟕)

= ( + 𝟏) =
𝟒 𝟔 𝟐𝟒
Therefore,
𝒑 𝒎 𝒏
𝟎, 𝒊𝒇 𝒂 > 3
𝟐
𝟏 𝒌 𝟏 𝒑(𝒑 + 𝟏)(𝟐𝒑 + 𝟕) 𝟏
𝛀 = 𝐥𝐢𝐦 𝒂 ⋅ ∑ ∑ ∑ 𝟐 𝟐
= 𝐥𝐢𝐦 𝒂 ⋅ = { , 𝒊𝒇 𝒂 = 𝟑
𝒑→∞ 𝒑 𝟐𝒌 − 𝟐𝒏𝒌 + 𝒏 𝒑→∞ 𝒑 𝟐𝟒 𝟏𝟐
𝒎=𝟏 𝒏=𝟏 𝒌=𝟏
+∞, 𝒊𝒇 𝒂 < 3
𝒌𝟐 (𝒏 − 𝒌)𝟐 𝒏𝟐 − 𝟐𝒏𝒌 + 𝒌𝟐
𝒂𝒌 = ⇒ 𝒂𝒏−𝒌 = =
𝟐𝒌𝟐 − 𝟐𝒏𝒌 + 𝒏𝟐 𝟐(𝒏 − 𝒌)𝟐 − 𝟐𝒏(𝒏 − 𝒌) + 𝒏𝟐 𝒏𝟐 − 𝟐𝒏𝒌 + 𝟐𝒌𝟐
𝒌𝟐 𝒏𝟐 − 𝟐𝒏𝒌 + 𝒌𝟐
𝒂𝒌 + 𝒂𝒏−𝒌 = + =
𝟐𝒌𝟐 − 𝟐𝒏𝒌 + 𝒏𝟐 𝒏𝟐 − 𝟐𝒏𝒌 + 𝟐𝒌𝟐
𝒌𝟐 (𝒏 − 𝒌 )𝟐
= +
𝟐𝒌𝟐 − 𝟐𝒏𝒌 + 𝒏𝟐 𝒏𝟐 − 𝟐𝒌𝒏 + 𝟐𝒌𝟐
𝒌 𝟐 + (𝒏 − 𝒌 )𝟐
𝒂𝒌 + 𝒂𝒏−𝒌 = =𝟏
𝒏𝟐 − 𝟐𝒌𝒏 + 𝟐𝒌𝟐
𝒏 𝒏
𝟏 𝒏
∑ 𝒂𝒌 = ∑ (𝒂𝒌 + 𝒂𝒏−𝒌 ) =
𝟐 𝟐
𝒌=𝟏 𝒌=𝟏
Hnece,
𝒌𝟐 𝒏+𝟏 𝒎(𝒎 + 𝟏) 𝒎
∑ ∑∑ 𝟐 = ∑ ∑ = ∑ ( + )=
𝟐𝒌 − 𝟐𝒏𝒌 + 𝒏𝟐 𝟐 𝟒 𝟐
𝒎=𝟏 𝒏=𝟏 𝒌=𝟏 𝒎=𝟏 𝒏=𝟏 𝒎=𝟏
𝒑 𝒑
𝟏 𝟏 𝟏 𝒑(𝒑 + 𝟏)(𝟐𝒑 + 𝟏) 𝟏 𝒑(𝒑 + 𝟏)
= ∑ 𝒎𝟐 + ∑ 𝒎 = ⋅ + ⋅ =
𝟒 𝟐 𝟒 𝟔 𝟐 𝟐
𝒎=𝟏 𝒎=𝟏
𝒑(𝒑 + 𝟏) 𝟐𝒑 + 𝟏 𝒑(𝒑 + 𝟏)(𝟐𝒑 + 𝟕)

= ( + 𝟏) =
𝟒 𝟔 𝟐𝟒
Therefore,
𝒑 𝒎 𝒏
𝟎, 𝒊𝒇 𝒂 > 3
𝟏 𝒌𝟐 𝟏 𝒑(𝒑 + 𝟏)(𝟐𝒑 + 𝟕) 𝟏
𝛀 = 𝐥𝐢𝐦 𝒂 ⋅ ∑ ∑ ∑ 𝟐 𝟐
= 𝐥𝐢𝐦 𝒂 ⋅ = { , 𝒊𝒇 𝒂 = 𝟑
𝒑→∞ 𝒑 𝟐𝒌 − 𝟐𝒏𝒌 + 𝒏 𝒑→∞ 𝒑 𝟐𝟒 𝟏𝟐
𝒎=𝟏 𝒏=𝟏 𝒌=𝟏
+∞, 𝒊𝒇 𝒂 < 3
www.ssmrmh.ro
Solution 3 by Soumitra Mandal-Chandar—Nagore-India
𝒌𝟐
Let 𝒂𝒌 = 𝟐𝒌𝟐 −𝟐𝒏𝒌+𝒏𝟐 . Therefore,
𝒌𝟐 (𝒏 − 𝒌 )𝟐
𝒂𝒌 + 𝒂𝒏−𝒌 = + =
𝟐𝒌𝟐 − 𝟐𝒏𝒌 + 𝒏𝟐 𝟐(𝒏 − 𝒌)𝟐 − 𝟐𝒏(𝒏 − 𝒌) + 𝒏𝟐
𝒌𝟐 (𝒏 − 𝒌 )𝟐
= + =
𝟐𝒌𝟐 − 𝟐𝒏𝒌 + 𝒏𝟐 (𝒏 − 𝒌)[𝟐(𝒏 − 𝒌) − 𝟐𝒏] + 𝒏𝟐
𝒌𝟐 (𝒏 − 𝒌 ) 𝟐 𝒌𝟐 (𝒏 − 𝒌 )𝟐
= + = + =
𝟐𝒌𝟐 − 𝟐𝒏𝒌 + 𝒏𝟐 −𝟐𝒌(𝒏 − 𝒌) + 𝒏𝟐 𝟐𝒌𝟐 − 𝟐𝒏𝒌 + 𝒏𝟐 𝒏𝟐 − 𝟐𝒌𝒏 + 𝟐𝒌𝟐
𝒌 𝟐 + (𝒏 − 𝒌 ) 𝟐
= =𝟏
𝒏𝟐 − 𝟐𝒌𝒏 + 𝟐𝒌𝟐
Hence,
𝒏−𝟏 𝒏−𝟏
𝟐 ∑ 𝒂𝒌 = ∑ (𝒂𝒌 + 𝒂𝒏−𝒌 ) = 𝒏 − 𝟏
𝒌=𝟏 𝒌=𝟏
and then
𝒏
𝒌𝟐 𝒏+𝟏
∑ 𝟐 𝟐
=
𝟐𝒌 − 𝟐𝒏𝒌 + 𝒏 𝟐
𝒌=𝟏
𝒌𝟐 𝒏+𝟏 𝒎(𝒎 + 𝟏) 𝒎
∑ ∑∑ 𝟐 𝟐
= ∑∑ = ∑( + )=
𝟐𝒌 − 𝟐𝒏𝒌 + 𝒏 𝟐 𝟒 𝟐
𝒎=𝟏 𝒏=𝟏 𝒌=𝟏 𝒎=𝟏 𝒏=𝟏 𝒎=𝟏
𝒑 𝒑
𝟏 𝟏 𝟏 𝒑(𝒑 + 𝟏)(𝟐𝒑 + 𝟏) 𝟏 𝒑(𝒑 + 𝟏)
= ∑ 𝒎𝟐 + ∑ 𝒎 = ⋅ + ⋅ =
𝟒 𝟐 𝟒 𝟔 𝟐 𝟐
𝒎=𝟏 𝒎=𝟏
𝒑(𝒑 + 𝟏) 𝟐𝒑 + 𝟏 𝒑(𝒑 + 𝟏)(𝟐𝒑 + 𝟕)

= ( + 𝟏) =
𝟒 𝟔 𝟐𝟒
Therefore,
𝒑 𝒎 𝒏
𝟎, 𝒊𝒇 𝒂 > 3
𝟏 𝒌𝟐 𝟏 𝒑(𝒑 + 𝟏)(𝟐𝒑 + 𝟕) 𝟏
𝛀 = 𝐥𝐢𝐦 𝒂 ⋅ ∑ ∑ ∑ 𝟐 𝟐
= 𝐥𝐢𝐦 𝒂 ⋅ = { , 𝒊𝒇 𝒂 = 𝟑
𝒑→∞ 𝒑 𝟐𝒌 − 𝟐𝒏𝒌 + 𝒏 𝒑→∞ 𝒑 𝟐𝟒 𝟏𝟐
𝒎=𝟏 𝒏=𝟏 𝒌=𝟏
+∞, 𝒊𝒇 𝒂 < 3

www.ssmrmh.ro
SP.398 If 𝒂, 𝒃 > 0, 𝑏 ≠ 1 solve for real numbers:
𝒏
𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒌) 𝒏+𝟏
∑ − =𝟎
𝟐𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒌)(𝒂 − 𝒌 + 𝒏) 𝟐
𝒌=𝟏

𝒙+𝐥𝐨𝐠 𝒃 (𝒂+𝒌)
Let 𝒂𝒌 = 𝟐𝒙+𝐥𝐨𝐠 . Hence,
𝒃 (𝒂+𝒌)(𝒂−𝒌+𝒏)
𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒌) 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒄 + 𝒏 − 𝒌)
𝒂𝒌 + 𝒂𝒏−𝒌 = + =𝟏
𝟐𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒌)(𝒂 − 𝒌 + 𝒏) 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏 − 𝒌)(𝒂 + 𝒌)
Hence,
𝒏−𝟏 𝒏−𝟏
𝟐 ∑ 𝒂𝒌 = ∑ (𝒂𝒌 + 𝒂𝒏−𝒌 ) = 𝒏 − 𝟏
𝒌=𝟏 𝒌=𝟏
and then
𝒏 𝒏−𝟏
𝒏−𝟏 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏)
∑ 𝒂𝒌 = ∑ 𝒂𝒌 + 𝒂𝒏 = +
𝟐 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 [𝒂(𝒂 + 𝒏)]
𝒌=𝟏 𝒌=𝟏
So, we get:
𝒏−𝟏 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏) 𝒏+𝟏 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏)
+ = ⇔ =𝟏⇔
𝟐 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 [𝒂(𝒂 + 𝒏)] 𝟐 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 [𝒂(𝒂 + 𝒏)]
𝟐𝒙 + 𝐥𝐨𝐠 𝒃 [𝒂(𝒂 + 𝒏)] = 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏)
𝟏
𝒙 = 𝐥𝐨𝐠 𝒃 ( )
𝒂
𝒏
𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒌)
𝑺=∑
𝟐𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒌)(𝒂 − 𝒌 + 𝒏)
𝒌=𝟏
𝒏−𝒌= 𝒑→ 𝒌 =𝒏−𝒑
𝒏−𝟏 𝒏
𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏 − 𝒑) 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏 − 𝒑)
𝑺=∑ =∑ +
𝟐𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏 − 𝒑)(𝒂 + 𝒑) 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏 − 𝒑)(𝒂 + 𝒑)
𝒑=𝟎 𝒑=𝟏
𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏) 𝒙 + 𝐥𝐨𝐠 𝒃 𝒂
+ −
𝟐𝒙 + 𝐥𝐨𝐠 𝒃 𝒂(𝒂 + 𝒏) 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 𝒂(𝒂 + 𝒏)

www.ssmrmh.ro
𝒏
𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏 − 𝒑)
⇒ 𝟐𝑺 = ∑ +
𝟐𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏 − 𝒑)(𝒂 + 𝒑)
𝒑=𝟎
+ − =
=𝒏+ −
𝒏 𝟏 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏) 𝒙 + 𝐥𝐨𝐠 𝒃 𝒂
𝑺= + ( − )
𝟐 𝟐 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 𝒂(𝒂 + 𝒏) 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 𝒂(𝒂 + 𝒏)
− =𝟏
So, we get:
𝒏−𝟏 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏) 𝒏+𝟏 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏)
+ = ⇔ =𝟏⇔
𝟐 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 [𝒂(𝒂 + 𝒏)] 𝟐 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 [𝒂(𝒂 + 𝒏)]
𝟐𝒙 + 𝐥𝐨𝐠 𝒃 [𝒂(𝒂 + 𝒏)] = 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏)
𝟏
𝒙 = 𝐥𝐨𝐠 𝒃 ( )
𝒂
𝒙+𝐥𝐨𝐠 𝒃 (𝒂+𝒌)
Let 𝒂𝒌 = 𝟐𝒙+𝐥𝐨𝐠 . Hence,
𝒃 (𝒂+𝒌)(𝒂−𝒌+𝒏)
𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒌) 𝒙 + 𝐥𝐨𝐠 𝒃 (𝒄 + 𝒏 − 𝒌)
𝒂𝒌 + 𝒂𝒏−𝒌 = + =𝟏
𝟐𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒌)(𝒂 − 𝒌 + 𝒏) 𝟐𝒙 + 𝐥𝐨𝐠 𝒃 (𝒂 + 𝒏 − 𝒌)(𝒂 + 𝒌)
Hence,
𝑺 = 𝒂𝟏 + 𝒂𝟐 + ⋯ + 𝒂𝒏
𝑺 + 𝒂𝟎 = 𝒂𝟎 + 𝒂𝟏 + ⋯ + 𝒂𝒏−𝟏 + 𝒂𝒏
𝑺 + 𝒂𝟎 = 𝒂𝒏 + 𝒂𝒏−𝟏 +. . . +𝒂𝟏 + 𝒂𝟎
𝒏+𝟏
So, we get: 𝟐(𝑺 + 𝒂𝟎 ) = 𝒏 + 𝟏 ⇒ 𝑺 = − 𝒂𝟎
𝟐
𝒏+𝟏 𝒏+𝟏
𝑺= − 𝒂𝟎 ⇒ 𝑺 − = 𝟎 ⇒ 𝒂𝟎 = 𝟎 ⇒ 𝒙 + 𝐥𝐨𝐠 𝒃 𝒂 = 𝟎
𝟐 𝟐
𝟏
𝒙 = 𝐥𝐨𝐠 𝒃 ( )
𝒂

www.ssmrmh.ro
SP.399 If 𝒎, 𝒏 > 0; 𝑥, 𝑦, 𝑧 ∈ (0,1) then:
𝟏 𝟏 𝟏 𝟗√𝟑
+ + ≥
(𝟏 − 𝒙𝟐 )(𝒎𝒚 + 𝒏𝒛) (𝟏 − 𝒚𝟐 )(𝒎𝒛 + 𝒏𝒙) (𝟏 − 𝒛𝟐 )(𝒎𝒙 + 𝒏𝒚) 𝟐
Proposed by D.M. Bătineţu-Giurgiu-Romania
𝟑
𝟐( 𝟐)
𝑨𝑴−𝑮𝑴 𝟐𝒕𝟐 + 𝟏 − 𝒕𝟐 + 𝟏 − 𝒕𝟐 𝟖
𝟐𝒕 𝟏 − 𝒕 ≤ ( ) =
𝟑 𝟐𝟕
𝟐
𝒕( 𝟏 − 𝒕 𝟐 ) ≤ ; 𝒕 ∈ (𝟎, 𝟏)
𝟑√𝟑
𝟏 𝒙 𝟑√𝟑 𝒙
∑ = ∑ ≥ ⋅ ∑ =
(𝟏 − 𝒙𝟐 )(𝒎𝒚 + 𝒏𝒛) 𝒙(𝟏 − 𝒙𝟐 )(𝒎𝒚 + 𝒏𝒛) 𝟑 𝒎𝒚 + 𝒏𝒛
𝟑√𝟑 𝒙𝟐 𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎 𝟑√𝟑 ( 𝒙 + 𝒚 + 𝒛) 𝟐

= ⋅∑ ≥ ⋅ ≥
𝟐 𝒎𝒙𝒚 + 𝒏𝒙𝒛 𝟐 (𝒎 + 𝒏)(𝒙 + 𝒚 + 𝒛)
𝒄𝒚𝒄
𝟑√𝟑 𝟑(𝒙𝒚 + 𝒚𝒛 + 𝒛𝒙) 𝟗√𝟑

≥ ⋅ =
𝟐 (𝒎 + 𝒏)(𝒙𝒚 + 𝒚𝒛 + 𝒛𝒙) 𝟐
𝟏
Equality holds for 𝒎 = 𝒏; 𝒙 = 𝒚 = 𝒛 = .
√𝟑
Solution 2 by George Florin Șerban-Romania

𝟏 𝟗√𝟑
∑ ≥ ; (∗)
(𝟏 − 𝒙𝟐 )(𝒎𝒚 + 𝒏𝒛) 𝟐(𝒎 + 𝒏)
𝒄𝒚𝒄
𝟐√𝟑 √𝟑 √𝟑 𝑨𝑮𝑴 𝟑
√𝟑 √𝟑
𝒙𝟐 + = 𝒙𝟐 + + ≥ 𝟑 ⋅ √ 𝒙𝟐 ⋅ ⋅ =𝟏
𝟗𝒙 𝟗𝒙 𝟗𝒙 𝟗𝒙 𝟗𝒙
𝟐√𝟑 𝟐√𝟑 𝟏 𝟗𝒙
⇒ 𝒙𝟐 + ≥ 𝟏; ∀𝒙 ∈ (𝟎, 𝟏) ⇒ 𝟏 − 𝒙𝟐 ≤ ⇒ ≥
𝟗𝒙 𝟗𝒙 𝟏 − 𝒙𝟐 𝟐√𝟑
𝟏 𝟗𝒙 𝟗 𝒙 (𝟏) 𝟗√𝟑
∑ 𝟐
≥ ∑ = ∑ ≥
(𝟏 − 𝒙 )(𝒎𝒚 + 𝒏𝒛) 𝟐√𝟑(𝒎𝒚 + 𝒏𝒛) 𝟐√𝟑 𝒄𝒚𝒄 𝒎𝒚 + 𝒏𝒛 𝟐(𝒎 + 𝒏)
𝒙 𝟑
( 𝟏) ⇔ ∑ ≥
𝒎𝒚 + 𝒏𝒛 𝒎 + 𝒏
𝒄𝒚𝒄
We have:

www.ssmrmh.ro
𝒙 𝒙𝟐 𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎 (𝒙 + 𝒚 + 𝒛)𝟐 ( 𝒙 + 𝒚 + 𝒛) 𝟐
∑ =∑ ≥ = ≥
𝒎𝒚 + 𝒏𝒛 𝒎𝒙𝒚 + 𝒏𝒙𝒛 𝒎 ∑ 𝒙𝒚 + 𝒏∑𝒙𝒛 (𝒎 + 𝒏)(𝒙𝒚 + 𝒚𝒛 + 𝒛𝒙)
𝟑
≥ , 𝐰𝐡𝐢𝐜𝐡 𝐢𝐬 𝐭𝐫𝐮𝐞 𝐟𝐫𝐨𝐦 (𝒙 + 𝒚 + 𝒛)𝟐 ≥ 𝟑(𝒙𝒚 + 𝒚𝒛 + 𝒛𝒙)
𝒎+𝒏
𝟏
√𝟑
Solution 3 by Soumitra Mandal-Chandar-Nagore-India
Let 𝒇(𝒙) = 𝒙 − 𝒙𝟑 , 𝒙 ≥ 𝟎, 𝒇′ (𝒙) = 𝟏 − 𝟑𝒙𝟐 , 𝒇′′ (𝒙) = −𝟔𝒙
𝟏 𝟏 𝟔
𝒇(𝒙) = 𝟎 ⇒ 𝒙 = . 𝐍𝐨𝐰 𝒇′′ ( )=− ; ∀𝒙 ≥ 𝟎
√𝟑 √𝟑 √𝟑
𝟏 𝟏
Hence, 𝒇 −attains maximum value at 𝒙 = then 𝒇 ( ) ≥ 𝒇(𝒙); ∀𝒙 ≥ 𝟎
√𝟑 √𝟑
𝟐
≥ 𝒇(𝒙); ∀𝒙 ≥ 𝟎
𝟑√𝟑
Hence,
𝟏 𝟏 𝒙 𝟑√𝟑 𝒙
∑ =∑ ⋅ ≥ ∑ =
(𝟏 − 𝒙𝟐 )(𝒎𝒚 + 𝒏𝒛) 𝟑
𝒙 − 𝒙 𝒎𝒚 + 𝒏𝒛 𝟐 𝒎𝒚 + 𝒏𝒛
𝟑√𝟑 𝒙𝟐 𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎 𝟑√𝟑 𝟏 ( 𝒙 + 𝒚 + 𝒛) 𝟐 𝟗√𝟑

= ∑ ≥ ⋅ ⋅ ≥
𝟐 𝒎𝒙𝒚 + 𝒏𝒙𝒛 𝟐 𝒎 + 𝒏 𝒙𝒚 + 𝒚𝒛 + 𝒛𝒙 𝟐(𝒎 + 𝒏)
𝒄𝒚𝒄
𝟏
√𝟑
SP.400 In 𝚫𝑨𝑩𝑪, 𝝀 > 0 the following relationship holds:
𝐬𝐢𝐧 𝑨 𝟑
∏( + 𝝀𝟐 ) ≥ 𝟗𝝀𝟐 ⋅ √
𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 𝟒
𝒄𝒚𝒄

For any positive real numbers 𝒙, 𝒚, 𝒛, 𝝀 holds inequality:
𝟑
(𝒙𝟐 + 𝝀𝟐 )(𝒚𝟐 + 𝝀𝟐 )(𝒛𝟐 + 𝝀𝟐 ) ≥ 𝝀𝟐 (𝒙 + 𝒚 + 𝒛)𝟐 ; (𝟏)
𝟒
Proof. We have:
(𝒙𝟐 + 𝝀𝟐 )(𝒚𝟐 + 𝝀𝟐 ) ≥ 𝝀𝟐 (𝒙 + 𝒚)𝟐 ⟺ (𝒙𝒚 − 𝝀𝟐 )𝟐 ≥ 𝟎
and

www.ssmrmh.ro
𝟐
𝟑 𝝀𝟐 𝒕𝟐
(𝒙 + 𝝀 𝒚 + 𝝀 ≥ 𝝀𝟐 ((𝒙 + 𝒚)𝟐 + 𝝀𝟐 ) ⟺ (𝒙𝒚 − ) + (𝒙 − 𝒚)𝟐 ≥ 𝟎
𝟐 𝟐 )( 𝟐 𝟐)
𝟒 𝟐 𝟒
Hence,
𝟑𝝀𝟐
(𝒙𝟐 + 𝝀𝟐 )(𝒚𝟐 + 𝝀𝟐 )(𝒛𝟐 + 𝝀𝟐 ) ≥ ((𝒙 + 𝒚)𝟐 + 𝝀𝟐 )(𝒛𝟐 + 𝝀𝟐 ) ≥
𝟒
𝟑 𝟐 𝟑
≥ 𝝀𝟐 ((𝒙 + 𝒚) + 𝒛) = 𝝀𝟐 (𝒙 + 𝒚 + 𝒛)𝟐
𝟒 𝟒
𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪
Let 𝒙 = √𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 , 𝒚 = √𝐬𝐢𝐧 𝑪 𝐬𝐢𝐧 𝑨 , 𝒛 = √𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 then
𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪
( + 𝝀𝟐 ) ( + 𝝀𝟐 ) ( + 𝝀𝟐 )
𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 𝐬𝐢𝐧 𝑪 𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩
𝟐
𝟑 𝟐 𝐬𝐢𝐧 𝑨
≥ 𝝀 (∑ √ ) ; (𝟐)
𝟒 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪
𝒄𝒚𝒄
We prove that:
𝐬𝐢𝐧 𝑨 𝟒 𝟒 𝟒 𝟒
∑√ ≥ 𝟑 ⋅ √ ; (𝟑) ⟺ ∑ 𝐬𝐢𝐧 𝑨 ≥ 𝟑 ⋅ √∏ 𝐬𝐢𝐧 𝑨 ⋅ √
𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 𝟑 𝟑
How
∑ 𝐬𝐢𝐧 𝑨 ≥ 𝟑 ⋅ 𝟑√∏ 𝐬𝐢𝐧 𝑨

Remains to prove that:

𝟒 𝟒 𝟑√𝟑
𝟑 ⋅ 𝟑√ ∏ 𝐬𝐢𝐧 𝑨 ≥ 𝟑 ⋅ √∏ 𝐬𝐢𝐧 𝑨 ⋅ √ ⟺ ∏ 𝐬𝐢𝐧 𝑨 ≤ (𝒕𝒓𝒖𝒆)
𝟑 𝟖
From (2),(3) it follows that:

𝐬𝐢𝐧 𝑨 𝟑 𝟒 𝟑
∏( + 𝝀𝟐 ) ≥ 𝝀𝟐 ⋅ 𝟗 ⋅ √ = 𝟗𝝀𝟐 ⋅ √
𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 𝟒 𝟑 𝟒
𝒄𝒚𝒄
𝟑
𝑯ӧ𝒍𝒅𝒆𝒓
𝐬𝐢𝐧 𝑨 𝟑 𝐬𝐢𝐧 𝑨 𝟑
∏( + 𝝀𝟐 ) ⏞
≥ ( √∏ + √(𝝀𝟐 )𝟑 )
𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪
𝟑
𝟏
= (𝟑 + 𝝀𝟐 ) ≥
√𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪
𝑨𝑴−𝑮𝑴 𝟑 𝑱𝒆𝒏𝒔𝒆𝒏
𝟑
⏞
≥ ( 𝟐
+𝝀 ) ⏞
≥
𝐬𝐢𝐧 𝑨 + 𝐬𝐢𝐧 𝑩 + 𝐬𝐢𝐧 𝑪

www.ssmrmh.ro
𝟑
𝟑 𝟑
𝟑 𝟐)
𝟐 𝟐
𝟏 𝟏 𝟐
( 𝝅+𝝀 =( +𝝀 ) =( + +𝝀 ) ≥
𝟑 𝐬𝐢𝐧 𝟑 √𝟑 √𝟑 √𝟑
𝑨𝑴−𝑮𝑴
𝟏 𝟏 𝟑
⏞
≥ 𝟐𝟕. . 𝝀𝟐 = 𝟗𝝀𝟐 ≥ 𝟗𝝀𝟐 . √
.
√𝟑 √𝟑 𝟒
𝐬𝐢𝐧 𝑨 𝟑
→ 𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆, ∏( + 𝝀𝟐 ) ≥ 𝟗𝝀𝟐 . √
𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 𝟒
𝒄𝒚𝒄
SP.401 Let 𝑰 be the incentre of triangle 𝑨𝑩𝑪 and let 𝑨′ , 𝑩′ and 𝑪′ be the
intersections of the rays 𝑨𝑰, 𝑩𝑰, and 𝑪𝑰 with the circumcircle of the triangle.
Prove that [𝑨′ 𝑩′ 𝑪′ ] ≥ [𝑨𝑩𝑪], where [∗] −represent the area.
𝑩 𝑪 𝝅 𝑨
We have: ∠𝑪′ 𝑨′ 𝑩′ = + 𝟐 = 𝟐 − 𝟐 . Using the law of sines, we have
𝟐
𝝅 𝑨 𝑪 𝑪 𝑩
𝑩′ 𝑪′ = 𝟐𝑹 ⋅ 𝐬𝐢𝐧 ( − ) = 𝟐𝑹 ⋅ 𝐜𝐨𝐬 ; 𝑨′ 𝑩′ = 𝟐𝑹 ⋅ 𝐜𝐨𝐬 ; 𝑪′ 𝑨′ = 𝟐𝑹 ⋅ 𝐜𝐨𝐬
𝟐 𝟐 𝟐 𝟐 𝟐
𝑨 𝑩 𝑪
𝑨′ 𝑩′ ⋅𝑩′ 𝑪′ ⋅𝑪′ 𝑨′ 𝟖𝑹𝟑 ⋅𝐜𝐨𝐬 𝐜𝐨𝐬 𝐜𝐨𝐬 𝑨 𝑩 𝑪
We know that: [𝑨′ 𝑩′ 𝑪′ ] = = 𝟐 𝟐 𝟐
= 𝟐𝑹𝟐 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐
𝟒𝑹 𝟒𝑹
𝑨 𝑩 𝑪
Now, we’ll prove that 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 ≥ 𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪. We have:
𝟏 𝟏 𝑪
𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 = (𝐜𝐨𝐬(𝑨 − 𝑩) − 𝐜𝐨𝐬(𝑨 + 𝑩)) ≤ (𝟏 + 𝐜𝐨𝐬 𝑪) = 𝐜𝐨𝐬 𝟐
𝟐 𝟐 𝟐
𝑪 𝑨 𝑩
So, 𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 ≤ 𝐜𝐨𝐬 𝟐 𝟐 and similarly 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 ≤ 𝐜𝐨𝐬 𝟐 𝟐 , 𝐬𝐢𝐧 𝑪 𝐬𝐢𝐧 𝑨 ≤ 𝐜𝐨𝐬 𝟐 𝟐
Multiplying up these inequalities, we get

𝑨 𝑩 𝑪 𝑨 𝑩 𝑪
𝐬𝐢𝐧𝟐 𝑨 𝐬𝐢𝐧𝟐 𝑩 𝐬𝐢𝐧𝟐 𝑪 ≤ 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 ⇔ 𝐜𝐨𝐬 𝐜𝐨𝐬 𝐜𝐨𝐬 ≥ 𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪
𝟐 𝟐 𝟐 𝟐 𝟐 𝟐
𝑨 𝑩 𝑪
Now, [𝑨′ 𝑩′ 𝑪′ ] = 𝟐𝑹𝟐 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 ≥ 𝟐𝑹𝟐 𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 =
𝒂 𝒃 𝒄 𝒂𝒃𝒄
= 𝟐𝑹𝟐 ( ) ( ) ( ) = = [𝑨𝑩𝑪], where 𝒂 = |𝑩𝑪|, 𝒃 = |𝑪𝑨|, 𝒄 = |𝑨𝑩|.
𝟐𝑹 𝟐𝑹 𝟐𝑹 𝟒𝑹
So, {𝑨′ 𝑩′ 𝑪′ ] ≥ [𝑨𝑩𝑪]

Equality holds if and only if triangle 𝑨𝑩𝑪 is equilateral.

www.ssmrmh.ro
̂′ = 𝟐𝑪
𝑨𝑩𝑨 ̂+𝑨
̂ ; 𝑩̂ ̂ +𝑪
′ 𝑨𝑪′ = 𝑩 ̂
𝟐𝑪+𝑨
In 𝚫𝑨𝑪𝑨′ by law of sines: 𝑨𝑨′ = 𝟐𝑹 𝐬𝐢𝐧 ( )
𝟐
𝑩+𝑪 𝑨
In 𝚫𝑨𝑩′𝑪′ by law of sines 𝑩′ 𝑪′ = 𝟐𝑹 𝐬𝐢𝐧 ( ) = 𝟐𝑹 𝐬𝐢𝐧 𝟐 and analogs.
𝟐
𝑨′ 𝑩′ ⋅ 𝑨′ 𝑪′ ⋅ 𝐬𝐢𝐧 𝑩̂
′ 𝑨′ 𝑪′ 𝟏 𝑪 𝑩 𝑨
[𝑨′ 𝑩′ 𝑪′ ] = = 𝟐𝑹 𝐬𝐢𝐧 ⋅ 𝟐𝑹 𝐬𝐢𝐧 ⋅ 𝐜𝐨𝐬 ⇒
𝟐 𝟐 𝟐 𝟐 𝟐
𝑨 𝑩 𝑪
[𝑨′ 𝑩′ 𝑪′ ] = 𝟐𝑹𝟐 ⋅ 𝐜𝐨𝐬 𝐜𝐨𝐬 𝐜𝐨𝐬 ; (𝟏)
𝟐 𝟐 𝟐
𝟐
But [𝑨𝑩𝑪] = 𝟐𝑹 ⋅ 𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 ; (𝟐)
From (1),(2) it follows that
𝑨 𝑩 𝑪 𝑨 𝑩 𝑪
[𝑨′ 𝑩′ 𝑪′ ] 𝟐𝑹𝟐 ⋅ 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐 𝐜𝐨𝐬 𝟐
= = =
[𝑨𝑩𝑪] 𝟐𝑹𝟐 𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪 𝑨 𝑨 𝑩 𝑩 𝑪 𝑪
𝐬𝐢𝐧 𝟐 𝐜𝐨𝐬 𝟐 𝐬𝐢𝐧 𝟐 𝐜𝐨𝐬 𝟐 𝐬𝐢𝐧 𝟐 𝐜𝐨𝐬 𝟐
𝟏
= ; (𝟑)
𝑨 𝑩 𝑪
𝟖 𝐬𝐢𝐧 𝐬𝐢𝐧 𝐬𝐢𝐧
𝟐 𝟐 𝟐
𝑨 𝑩 𝑪 𝒓
But 𝐬𝐢𝐧 𝟐 𝐬𝐢𝐧 𝟐 𝐬𝐢𝐧 𝟐 = 𝟒𝑹 ; (𝟒). From (3),(4) we have
[𝑨′ 𝑩′ 𝑪′ ] 𝑹
[𝑨𝑩𝑪]
= 𝟐𝒓 ; (𝟓) and 𝑹 ≥ 𝟐𝒓 (𝑬𝒖𝒍𝒆𝒓); (𝟔)
From (5),(6) we get [𝑨′ 𝑩′ 𝑪′ ] ≥ [𝑨𝑩𝑪]

SP.402 Let 𝒎𝒂 , 𝒎𝒃𝒎𝒄 be the lengths of the medians of a triangle 𝑨𝑩𝑪 with
inradius 𝒓 and circumradius 𝑹. Prove that:
𝟖𝒓𝟐 𝐬𝐢𝐧𝟐 𝑩 + 𝐬𝐢𝐧𝟐 𝑪 𝟏
≤ ∑ ≤
𝑹𝟒 𝒎𝟐𝒂 𝟐𝒓𝟐
𝒄𝒚𝒄

Let 𝒂 = |𝑩𝑪|, 𝒃 = |𝑪𝑨|, 𝒄 = |𝑨𝑩|. We know that 𝒔𝒂 ≥ 𝒉𝒂, with 𝒔𝒂 =
𝟐𝒃𝒄 𝒃𝒄 𝟐𝒎 𝟏 𝒎𝟐 𝒎𝒂 𝒃𝟐 +𝒄𝟐 𝟒𝑹
𝒎𝒂 (𝒔𝒚𝒎𝒎𝒆𝒅𝒊𝒂𝒏) and 𝒉𝒂 = 𝟐𝑹 .So, 𝒃𝟐 +𝒄𝒂𝟐 ≥ 𝟐𝑹 ⇔ 𝒃𝟐 +𝒄
𝒂
𝟐 ≥ ⇔ ≤𝒎 .
𝒃𝟐 +𝒄𝟐 𝟒𝑹 𝒎𝟐𝒂 𝒂

www.ssmrmh.ro
𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
Also, we know that 𝒎𝒂 ≥ 𝒉𝒂 (𝒂𝒏𝒅 𝒂𝒏𝒂𝒍𝒐𝒈𝒔) ⇔ 𝒎 ≤ 𝒉 and 𝒉 + 𝒉 + 𝒉 = 𝒓.
𝒂 𝒂 𝒂 𝒃 𝒄
𝒃𝟐 + 𝒄𝟐 𝟏 𝟏 𝟒𝑹/𝟏 𝟒𝑹
∑ 𝟐
≤ 𝟒𝑹 ∑ ≤ 𝟒𝑹 ∑ = =
𝒎𝒂 𝒎𝒂 𝒉𝒂 𝒓 𝒓
Using the law of the sines, we get:

𝐬𝐢𝐧𝟐 𝑩 + 𝐬𝐢𝐧𝟐 𝑪 𝟒 𝑬𝒖𝒍𝒆𝒓 𝟏
∑ ≤ ≤
𝒎𝟐𝒂 𝑹𝒓 𝟐𝒓𝟐
𝒄𝒚𝒄
For the left inequality, we have:

𝒃𝟐 + 𝒄𝟐 𝒄𝟐 𝒂𝟐 𝒃𝟐 𝒃𝟐 𝒄𝟐 𝒂𝟐 𝑪𝑩𝑺
∑ = ( 𝟐 + 𝟐 + 𝟐) + ( 𝟐 + 𝟐 + 𝟐) ≥
𝒎𝟐𝒂 𝒎𝒂 𝒎𝒃 𝒎𝒄 𝒎𝒂 𝒎𝒃 𝒎𝒄
𝒄𝒚𝒄
(𝒄 + 𝒂 + 𝒃)𝟐 (𝒃 + 𝒄 + 𝒂)𝟐 𝟖𝒔𝟐

≥ + = ; 𝒂 + 𝒃 + 𝒄 = 𝟐𝒔
𝒎𝟐𝒂 + 𝒎𝟐𝒃 + 𝒎𝟐𝒄 𝒎𝟐𝒂 + 𝒎𝟐𝒃 + 𝒎𝟐𝒄 𝒎𝟐𝒂 + 𝒎𝟐𝒃 + 𝒎𝟐𝒄
𝟑
We know that: 𝒎𝟐𝒂 + 𝒎𝟐𝒃 + 𝒎𝟐𝒄 = 𝟒 (𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ) and 𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ≤ 𝟗𝑹𝟐 .
𝟐𝟕
So, 𝒎𝟐𝒂 + 𝒎𝟐𝒃 + 𝒎𝟐𝒄 ≤ 𝑹𝟐 . Hence,
𝟒
𝒃𝟐 + 𝒄𝟐 𝟖𝒔𝟐 𝑴𝒊𝒕𝒓𝒊𝒏𝒐𝒗𝒊𝒄 𝟑𝟐𝒓𝟐

∑ ≥ ≥
𝒎𝟐𝒂 𝟐𝟕 𝟐 𝑹𝟐
𝒄𝒚𝒄
𝟒 𝑹
Using Law of sines, it follows that
𝐬𝐢𝐧𝟐 𝑩 + 𝐬𝐢𝐧𝟐 𝑪 𝟖𝒓𝟐
∑ ≥ 𝟒
𝒎𝟐𝒂 𝑹
𝒄𝒚𝒄
Therefore,
𝟖𝒓𝟐 𝐬𝐢𝐧𝟐 𝑩 + 𝐬𝐢𝐧𝟐 𝑪 𝟏
𝟒
≤ ∑ 𝟐
≤ 𝟐
𝑹 𝒎𝒂 𝟐𝒓
𝒄𝒚𝒄
Solution 2 by Avishek Mitra-West Bengal-India

𝒃𝟐 +𝒄𝟐
𝟐 𝟐 𝟐 𝟐 𝒎𝒂 ≥
𝐬𝐢𝐧 𝑩 + 𝐬𝐢𝐧 𝑪 𝒃 +𝒄 𝟏 𝟒𝑹 (𝒃𝟐 + 𝒄𝟐 ) 𝟏𝟔𝑹𝟐
∑ 𝟐
=∑ 𝟐
⋅ 𝟐 ≤ ∑ ⋅ =
𝒎𝒂 𝟒𝑹 𝒎𝒂 𝟒𝑹𝟐 (𝒃𝟐 + 𝒄𝟐 )𝟐
𝟏 𝟏 𝟐 𝟒𝒔 𝟐 𝑬𝒖𝒍𝒆𝒓 𝟏
= 𝟒∑ = 𝟒 ∑ = ∑ 𝒂 = = ≤
𝒃𝟐 + 𝒄𝟐 𝟐𝒃𝒄 𝒂𝒃𝒄 𝟒𝑹𝒓𝒔 𝑹𝒓 𝟐𝒓𝟐

www.ssmrmh.ro
𝟐 𝟐
(𝐬𝐢𝐧 𝑩 + 𝐬𝐢𝐧 𝑪)𝟐
𝐬𝐢𝐧 𝑩 + 𝐬𝐢𝐧 𝑪 (𝑷𝒐𝒘𝒆𝒓−𝑴𝒆𝒂𝒏) 𝑳𝒆𝒊𝒃𝒏𝒊𝒛
∑ ≥ ∑ 𝟐 ≥
𝒎𝟐𝒂 𝒎𝟐𝒂
𝟏 (∑(𝐬𝐢𝐧 𝑩 + 𝐬𝐢𝐧 𝑪))𝟐 𝟏 (𝟐 ∑ 𝐬𝐢𝐧 𝑨)𝟐 𝑳𝒆𝒊𝒃𝒏𝒊𝒛

≥ ⋅ = ⋅ ≥
𝟐 ∑ 𝒎𝟐𝒂 𝟐 𝟑
∑ 𝟐
𝟒 𝒂
𝒂 𝟐
𝟏 (𝟐 ∑ 𝟐𝑹) 𝟏 𝟒 𝑴𝒊𝒕𝒓𝒊𝒏𝒐𝒗𝒊𝒄 𝟖𝒓𝟐
≥ ⋅ = ⋅ ⋅ 𝟒𝒔𝟐 ≥
𝟐 𝟑 ⋅ 𝟗𝑹𝟐 𝟐𝑹𝟒 𝟐𝟕 𝑹𝟒
𝟒
SP.403 In 𝚫𝑨𝑩𝑪, 𝝀 > 0, 𝑛 ∈ ℕ the following relationship holds:
𝐬𝐢𝐧𝟐𝒏 𝑨
∏ ( 𝟐𝒏+𝟒 + 𝝀𝟐 ) ≥ 𝟏𝟐𝝀𝟐
𝐬𝐢𝐧 𝑩
𝒄𝒚𝒄

For any positive real numbers 𝒙, 𝒚, 𝒛, 𝝀 holds inequality:
𝟑 𝟐
(𝒙𝟐 + 𝝀𝟐 )(𝒚𝟐 + 𝝀𝟐 )(𝒛𝟐 + 𝝀𝟐 ) ≥ 𝝀 (𝒙 + 𝒚 + 𝒛)𝟐 ; (𝟏)
𝟒
Proof. We have:
(𝒙𝟐 + 𝝀𝟐 )(𝒚𝟐 + 𝝀𝟐 ) ≥ 𝝀𝟐 (𝒙 + 𝒚)𝟐 ⟺ (𝒙𝒚 − 𝝀𝟐 )𝟐 ≥ 𝟎
and
𝟐
𝟐 𝟐 )( 𝟐 𝟐)
𝟑 𝝀𝟐 𝒕𝟐
(𝒙 + 𝝀 𝒚 +𝝀 ≥ 𝝀𝟐 ((𝒙 + 𝒚)𝟐 + 𝝀𝟐 ) ⟺ (𝒙𝒚 − ) + (𝒙 − 𝒚)𝟐 ≥ 𝟎
𝟒 𝟐 𝟒
Hence,
𝟑𝝀𝟐
(𝒙𝟐 + 𝝀𝟐 )(𝒚𝟐 + 𝝀𝟐 )(𝒛𝟐 + 𝝀𝟐 ) ≥ ((𝒙 + 𝒚)𝟐 + 𝝀𝟐 )(𝒛𝟐 + 𝝀𝟐 ) ≥
𝟒
𝟑 𝟐 𝟑
≥ 𝝀𝟐 ((𝒙 + 𝒚) + 𝒛) = 𝝀𝟐 (𝒙 + 𝒚 + 𝒛)𝟐
𝟒 𝟒
𝐬𝐢𝐧𝒏 𝑨 𝐬𝐢𝐧𝒏 𝑩 𝐬𝐢𝐧𝒏+𝟏 𝑪
Let 𝒙 = 𝐬𝐢𝐧𝒏+𝟐 𝑩 , 𝒚 = 𝐬𝐢𝐧𝒏+𝟐 𝑪 , 𝒛 = 𝐬𝐢𝐧𝒏+𝟐 𝑨 then
𝟐
𝐬𝐢𝐧𝟐𝒏 𝑨 𝟐
𝟑 𝟐 𝐬𝐢𝐧𝒏 𝑨
∏ ( 𝟐𝒏+𝟒 + 𝝀 ) ≥ 𝝀 (∑ ) ; ( 𝟐)
𝐬𝐢𝐧 𝑩 𝟒 𝐬𝐢𝐧𝒏+𝟐 𝑩

www.ssmrmh.ro
Remains to prove that:
𝐬𝐢𝐧𝒏 𝑨
∑ ≥ 𝟒; (𝟑)
𝐬𝐢𝐧𝒏+𝟐 𝑩
𝒄𝒚𝒄
From AM-GM inequality, we have:
𝐬𝐢𝐧𝒏 𝑨 𝟑 𝟏 (∗) 𝟑 𝒓𝒔 𝟑√𝟑

∑ ≥ 𝟑 ⋅ √ ≥ = 𝟒, (∗) ⟺ ∏ 𝐬𝐢𝐧 𝑨 = ≤
𝐬𝐢𝐧𝒏+𝟐 𝑩 (∏ 𝐬𝐢𝐧 𝑨)𝟐 𝟑
𝟐 𝟐𝑹 𝟐 𝟖
√(𝟑√𝟑)
𝟖

𝟐
𝟐𝒏 𝒏
𝐬𝐢𝐧 𝑨 𝟑 𝐬𝐢𝐧 𝑨
∏( 𝟐𝒏+𝟒
+ 𝝀𝟐 ) ≥ 𝝀𝟐 (∑ ) ≥ 𝟏𝟐𝝀𝟐
𝐬𝐢𝐧 𝑩 𝟒 𝐬𝐢𝐧𝒏+𝟐 𝑩

𝟑
𝑯ӧ𝒍𝒅𝒆𝒓
𝐬𝐢𝐧𝟐𝒏 𝑨 𝟑 𝐬𝐢𝐧𝟐𝒏 𝑨 𝟑
∏ ( 𝟐𝒏+𝟒 + 𝝀𝟐 ) ⏞
≥ ( √∏ + √(𝝀𝟐 )𝟑 )
𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧𝟐𝒏+𝟒 𝑩
𝟑
𝟏
=( 𝟒 + 𝝀𝟐 ) ≥
𝟑
√𝐬𝐢𝐧 𝑨 𝐬𝐢𝐧 𝑩 𝐬𝐢𝐧 𝑪
𝟒 𝟑
𝑨𝑴−𝑮𝑴 𝟒 𝟑 𝑱𝒆𝒏𝒔𝒆𝒏 𝟒 𝟑
𝟑 𝟑 𝟐
⏞ [( ) + 𝝀𝟐 ] ⏞ [( 𝟐 𝟐
≥
𝐬𝐢𝐧 𝑨 + 𝐬𝐢𝐧 𝑩 + 𝐬𝐢𝐧 𝑪
≥ 𝝅) + 𝝀 ] = [(√𝟑) + 𝝀 ]
𝟑 𝐬𝐢𝐧 𝟑
𝟑
𝟑 𝑨𝑴−𝑮𝑴
𝟖 𝟖 𝟖 𝟖
𝟑 𝟔𝟒𝝀𝟐
= ( + + 𝝀𝟐 ) ⏞
≥ (𝟑. √ . . 𝝀𝟐 ) = ≥ 𝟏𝟐𝝀𝟐
𝟗 𝟗 𝟗 𝟗 𝟑
𝐬𝐢𝐧𝟐𝒏 𝑨
𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆, ∏( + 𝝀𝟐 ) ≥ 𝟏𝟐𝝀𝟐
𝐬𝐢𝐧𝟐𝒏+𝟒 𝑩
𝒄𝒚𝒄
SP.404 In 𝚫𝑨𝑩𝑪 the following relationship holds:
𝟑(𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ) + 𝟒(𝒉𝟐𝒂 + 𝒉𝟐𝒃 + 𝒉𝟐𝒄 ) ≥ 𝟐𝟒√𝟑𝑭

𝟐(𝒃𝟐 + 𝒄𝟐 ) − 𝒂𝟐 𝒂𝟐
𝟐𝒉𝟐𝒂 − 𝟐𝒎𝟐𝒂 = 𝟐𝒉𝟐𝒂 −𝟐⋅ 𝟐 𝟐 𝟐
= 𝟐𝒉𝒂 − 𝒃 − 𝒄 +
𝟒 𝟐
www.ssmrmh.ro
Hence,
𝒂𝟐 𝟑 𝑨𝑴−𝑮𝑴
𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 + 𝟐𝒉𝟐𝒂 − 𝟐𝒎𝟐𝒂 = 𝟐𝒉𝟐𝒂 + 𝒂𝟐 + = 𝟐𝒉𝟐𝒂 + 𝒂𝟐 ≥
𝟐 𝟐
𝟑𝒂𝟐
≥ 𝟐√𝟐𝒉𝟐𝒂 ⋅ = 𝟐√𝟑𝒂𝟐 𝒉𝟐𝒂 = 𝟐√𝟐(𝟐𝑭)𝟐 = 𝟒√𝟑𝑭
𝟐
𝟑(𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ) + 𝟐(𝒉𝟐𝒂 + 𝒉𝟐𝒃 + 𝒉𝟐𝒄 ) − 𝟐(𝒎𝟐𝒂 + 𝒎𝟐𝒃 + 𝒎𝟐𝒄 ) ≥ 𝟏𝟐√𝟑𝑭

Thus,
𝟑
𝟑(𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ) + 𝟐(𝒉𝟐𝒂 + 𝒉𝟐𝒃 + 𝒉𝟐𝒄 ) − 𝟐 ⋅ (𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ) ≥ 𝟏𝟐√𝟑𝑭
𝟒
𝟑 𝟐
(𝒂 + 𝒃𝟐 + 𝒄𝟐 ) + 𝟐(𝒉𝟐𝒂 + 𝒉𝟐𝒃 + 𝒉𝟐𝒄 ) ≥ 𝟏𝟐√𝟑𝑭
𝟐
Therefore,
𝟑(𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ) + 𝟒(𝒉𝟐𝒂 + 𝒉𝟐𝒃 + 𝒉𝟐𝒄 ) ≥ 𝟐𝟒√𝟑𝑭
𝟏𝟔𝑭𝟐 𝑨𝑮𝑴
We have 𝟑𝒂𝟐 + 𝟒𝒉𝟐𝒂 = 𝟑𝒂𝟐 + ≥ 𝟖𝑭√𝟑 (and analogs for 𝒃 and 𝒄)
𝒂𝟐
Summing these relationships, we find:

𝟑(𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 ) + 𝟒(𝒉𝟐𝒂 + 𝒉𝟐𝒃 + 𝒉𝟐𝒄 ) ≥ 𝟐𝟒√𝟑𝑭.
SP.405 If 𝒂, 𝒃, 𝒄 > 0 then:

√𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 + √𝒃𝟐 + 𝟓𝒃𝒄 + 𝟕𝒄𝟐 + √𝒄𝟐 + 𝟓𝒄𝒂 + 𝟕𝒂𝟐 ≥ √𝟏𝟑(𝒂 + 𝒃 + 𝒄)
𝟑(𝒂 − 𝒃)𝟐 + (𝟕𝒂 + 𝟏𝟗𝒃)𝟐 = 𝟑(𝒂𝟐 − 𝟐𝒂𝒃 + 𝒃𝟐 ) + 𝟒𝟗𝒂𝟐 + 𝟑𝟔𝟏𝒃𝟐 + 𝟐𝟔𝟔𝒂𝒃 =
= 𝟓𝟐𝒂𝟐 + 𝟐𝟔𝟎𝒂𝒃 + 𝟑𝟔𝟒𝒃𝟐 = 𝟓𝟐(𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 )
Hence,
𝟏 𝟏
𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 = (𝟑(𝒂 − 𝒃)𝟐 + (𝟕𝒂 + 𝟏𝟗𝒃)𝟐 ) ≥ (𝟕𝒂 + 𝟏𝟗𝒃)𝟐
𝟓𝟐 𝟓𝟐
So, we have:

www.ssmrmh.ro
𝟏
√𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 ≥ (𝟕𝒂 + 𝟏𝟗𝒃)
√𝟓𝟐
Therefore,
𝟏
∑ √𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 ≥ (𝟕 ∑ 𝒂 + 𝟏𝟗 ∑ 𝒃) =
𝒄𝒚𝒄
𝟐√𝟏𝟑 𝒄𝒚𝒄 𝒄𝒚𝒄
𝟏 𝟐𝟔
= (𝟕(𝒂 + 𝒃 + 𝒄) + 𝟏𝟗(𝒂 + 𝒃 + 𝒄)) = (𝒂 + 𝒃 + 𝒄) = √𝟏𝟑(𝒂 + 𝒃 + 𝒄)
𝟐√𝟏𝟑 𝟐√𝟏𝟑
𝑨𝑮𝑴 𝟏𝟑 𝟐𝟔 𝟕 𝟏𝟗
∑ √𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 ≥ ∑ √𝟏𝟑 √𝒂𝟕 ⋅ 𝒃𝟏𝟗 = √𝟏𝟑 ∑ √𝒂𝟕 ⋅ 𝒃𝟏𝟗 = √𝟏𝟑 ∑ 𝒂𝟐𝟔 ⋅ 𝒃𝟐𝟔
𝟏 𝟏
𝒏 𝒏 𝒑 𝒏 𝒒
𝒑 𝒒 𝟏 𝟏
𝐁𝐲 𝐇𝐨𝐥𝐝𝐞𝐫: ∑ 𝒂𝒊 𝒃𝒊 ≤ (∑ 𝒂𝒊 ) ⋅ (∑ 𝒃𝒊 ) , 𝐰𝐡𝐞𝐫𝐞 + = 𝟏.
𝒑 𝒒
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏
𝟕 𝟏𝟗
𝟐𝟔 𝟐𝟔 𝟐𝟔 𝟐𝟔
𝟕 𝟏𝟗 𝟕 𝟕 𝟏𝟗 𝟏𝟗
𝐅𝐨𝐫 𝒏 = 𝟑: ∑ 𝒂𝟐𝟔 ⋅ 𝒃𝟐𝟔 ≥ (∑ (𝒂𝟐𝟔 ) ) ⋅ (∑ (𝒃𝟐𝟔 ) ) =
𝟕 𝟏𝟗
𝟐𝟔 𝟐𝟔
= (∑ 𝒂) ⋅ (∑ 𝒃) = 𝒂 + 𝒃 + 𝒄 ⇒ ∑ √𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 ≥ √𝟏𝟑(𝒂 + 𝒃 + 𝒄).


𝟐√𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 = √(𝟐𝒂 + 𝟓𝒃)𝟐 + 𝟑𝒃𝟐
𝟏 𝟐𝒂 + 𝟓𝒃 𝟐 𝟐𝒂 + 𝟓𝒃 𝟐
= √[( ) + ⋯+ ( ) + 𝒃𝟐 + 𝒃𝟐 + 𝒃𝟐 ] (𝟏
⏟+ ⋯ + 𝟏) ≥
√𝟓𝟐 ⏟ 𝟕 𝟕 𝟓𝟐 𝒕𝒊𝒎𝒆𝒔
𝟒𝟗 𝒕𝒊𝒎𝒆𝒔
𝑪𝑩𝑺
√𝟏𝟑 𝟐𝒂 + 𝟓𝒃 𝟐𝒂 + 𝟓𝒃 √𝟏𝟑
⏞
≥ (( )+⋯+( ) + 𝒃 + 𝒃 + 𝒃) = (𝟕𝒂 + 𝟏𝟗𝒃)
𝟐𝟔 ⏟ 𝟕 𝟕 𝟏𝟑
𝟒𝟗 𝒕𝒊𝒎𝒆𝒔
√𝟏𝟑
→ √𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 ≥ (𝟕𝒂 + 𝟏𝟗𝒃) (𝑨𝒏𝒅 𝒂𝒏𝒂𝒍𝒐𝒈𝒔)
𝟐𝟔
www.ssmrmh.ro
√𝟏𝟑 √𝟏𝟑
→ ∑ √𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 ≥ ∑(𝟕𝒂 + 𝟏𝟗𝒃) = . 𝟐𝟔 ∑ 𝒂 = √𝟏𝟑(𝒂 + 𝒃 + 𝒄)
𝟐𝟔 𝟐𝟔
𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆, ∑ √𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 ≥ √𝟏𝟑(𝒂 + 𝒃 + 𝒄)

𝒄𝒚𝒄
∑ √𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 =

𝒄𝒚𝒄
𝟐 𝟐
𝑴𝒊𝒏𝒌𝒐𝒘𝒔𝒌𝒊
𝟏 𝟏
= ∑ √(𝟐𝒂 + 𝟓𝒃)𝟐 + 𝟑𝒃𝟐 ⏞
≥ √(∑(𝟐𝒂 + 𝟓𝒃)) + 𝟑 (∑ 𝒃) =
𝟐 𝟐
𝟐 𝟐
𝟏
= √𝟒𝟗 (∑ 𝒂) + 𝟑 (∑ 𝒂) = √𝟏𝟑 ∑ 𝒂.
𝟐
𝐓𝐡𝐞𝐫𝐞𝐟𝐨𝐫𝐞, ∑ √𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 ≥ √𝟏𝟑(𝒂 + 𝒃 + 𝒄)

𝒄𝒚𝒄

Solution 5 by Angel Plaza-Gran Canaria-Spain
The proposed inequality is equivalent to
√𝒃𝟐 + 𝟓𝒃𝒂 + 𝟕𝒂𝟐 + √𝒄𝟐 + 𝟓𝒄𝒃 + 𝟕𝒃𝟐 + √𝒂𝟐 + 𝟓𝒂𝒄 + 𝟕𝒄𝟐 ≥ √𝟏𝟑(𝒂 + 𝒃 + 𝒄)
Or, after summing up both of them, to
𝑺[𝒂, 𝒃, 𝒄] + 𝑺[𝒃, 𝒂, 𝒄] ≥ 𝟐√𝟏𝟑(𝒂 + 𝒃 + 𝒄),
where 𝑺[𝒂, 𝒃, 𝒄] and 𝑺[𝒃, 𝒂, 𝒄] respectively denote the left-hand side of both inequalities.
We will prove that
√𝒂𝟐 + 𝟓𝒂𝒃 + 𝟕𝒃𝟐 + √𝒃𝟐 + 𝟓𝒃𝒂 + 𝟕𝒂𝟐 ≥ √𝟏𝟑(𝒂 + 𝒃)

Last inequality is homogeneous, so we may assume that 𝒂 + 𝒃 = 𝟏, and the inequality
becomes 𝒇(𝒂) + 𝒇(𝟏 − 𝒂) ≥ √𝟏𝟑, where 𝒇(𝒙) = √𝒙𝟐 + 𝟓𝒙(𝟏 − 𝒙) + 𝟕(𝟏 − 𝒙)𝟐 .
𝟑
Function 𝒇(𝒙) is convex because 𝒇′′ (𝒙) = > 𝟎, ∀𝒙 ∈ [𝟎, 𝟏].
𝟑
𝟒√(𝟑𝒙𝟐 −𝟗𝒙+𝟕)

www.ssmrmh.ro
𝟏
Therefore, by Jensen’s inequality 𝒇(𝒂) + 𝒇(𝟏 − 𝒂) ≥ 𝟐𝒇 (𝟐) = √𝟏𝟑, and the problem is
done. Equality holds for 𝒂 = 𝒃 = 𝒄.
UNDERGRADUATE PROBLEMS
UP.391
𝒏
(𝟐𝒊 − 𝟏)𝒙
𝒙𝒏 = ∑ 𝐬𝐢𝐧 ,𝒙 > 0
𝒏𝟐
𝒊=𝟏
Find:
𝟏
−𝟏 𝒙𝒏
𝒏 𝒏
𝟏+𝒙
𝛀 = 𝐥𝐢𝐦 𝐥𝐢𝐦 ((∑ 𝒂𝒊 𝒏 ) ⋅ (∑ 𝒂𝒊 ) ) , 𝒂𝒊 > 0, 𝑖 = ̅̅̅̅̅
𝟏, 𝒏
𝒙→𝟎 𝒏→∞
𝒊=𝟏 𝒊=𝟏
( )
For 𝒂 > 0. We prove that:
𝒏
(𝟐𝒊 − 𝟏)𝒂
𝐥𝐢𝐦 𝒙𝒏 = 𝐥𝐢𝐦 (∑ 𝐬𝐢𝐧 )=𝒂
𝒏→∞ 𝒏→∞ 𝒏𝟐
𝒊=𝟏
𝒙𝟑
Using the well-known inequality: 𝒙 − < 𝐬𝐢𝐧 𝒙 < 𝑥, ∀𝑥 > 0 ⇒
𝟔
𝒂 𝟏 𝒂𝟑 𝒂 𝒂
𝟐
− ⋅ 𝟔 < 𝐬𝐢𝐧 𝟐 < 𝟐
𝒏 𝟔 𝒏 𝒏 𝒏
𝟑𝒂 𝟏 𝟑𝟑 𝒂𝟑 𝟑𝒂 𝟑𝒂
𝟐
− ⋅ 𝟔 < 𝐬𝐢𝐧 𝟐 < 𝟐
𝒏 𝟔 𝒏 𝒏 𝒏
………………………………
(𝟐𝒏 − 𝟏)𝒂 𝟏 (𝟐𝒏 − 𝟏)𝟑 𝒂𝟑 (𝟐𝒏 − 𝟏)𝒂 (𝟐𝒏 − 𝟏)𝒂
𝟐
− ⋅ 𝟔
< 𝐬𝐢𝐧 <
𝒏 𝟔 𝒏 𝒏𝟐 𝒏𝟐
Summing, we get:
𝒏 𝒏 𝒏 𝒏
𝒂 𝒂𝟑 𝟑
(𝟐𝒊 − 𝟏)𝒂 𝒂
⋅ ∑(𝟐𝒊 − 𝟏) − ⋅ ∑( 𝟐𝒊 − 𝟏 ) < ∑ 𝐬𝐢𝐧 < ⋅ ∑(𝟐𝒊 − 𝟏)
𝒏𝟐 𝒏𝟔 𝒏𝟐 𝒏𝟐
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝒊=𝟏

www.ssmrmh.ro
Let us denote:
𝒏 𝒏
𝒂 𝒂𝟑
𝒂𝒏 = 𝟐 ⋅ ∑(𝟐𝒊 − 𝟏) , 𝒃𝒏 = 𝒙𝒏 − 𝟔 ⋅ ∑(𝟐𝒊 − 𝟏)𝟑
𝒏 𝒏
𝒊=𝟏 𝒊=𝟏
𝒏
𝒂𝟑 𝒂𝟑
⇒ 𝒂𝒏 = 𝒂 ⇒ 𝒃𝒏 = 𝒂 − 𝟔 ⋅ ∑(𝟐𝒊 − 𝟏)𝟑 = 𝒂 − 𝟒
𝒏 𝒏 (𝟐𝒏𝟐 − 𝟏)
𝒊=𝟏
So, 𝐥𝐢𝐦 𝒂𝒏 = 𝐥𝐢𝐦 𝒃𝒏 = 𝒂 then,

𝒏→∞ 𝒏→∞
𝒏
(𝟐𝒊 − 𝟏)𝒙
𝐥𝐢𝐦 𝒙𝒏 = 𝐥𝐢𝐦 (∑ 𝐬𝐢𝐧 )=𝒙
𝒏→∞ 𝒏→∞ 𝒏𝟐
𝒊=𝟏
𝟏 𝟏
−𝟏 𝒙𝒏 −𝟏 𝒙
𝒏 𝒏 𝒏 𝒏
𝟏+𝒙𝒏
𝛀 = 𝐥𝐢𝐦 𝐥𝐢𝐦 ((∑ 𝒂𝒊 ) ⋅ (∑ 𝒂𝒊 ) ) = 𝐥𝐢𝐦 ((∑ 𝒂𝟏+𝒙
𝒊 ) ⋅ (∑ 𝒂𝒊 ) )
𝒙→𝟎 𝒏→∞ 𝒙→𝟎
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝒊=𝟏
( )
Let be the function
𝟏
−𝟏 𝒙
𝒏 𝒏
𝒇(𝒙) = ((∑ 𝒂𝟏+𝒙

𝒊 ) ⋅ (∑ 𝒂𝒊 ) )
𝒊=𝟏 𝒊=𝟏
𝒏 𝒏
𝟏
𝒇(𝒙) = 𝒆𝐥𝐨𝐠 𝒈(𝒙) ; 𝒈(𝒙) = (𝐥𝐨𝐠 (∑ 𝒂𝒙+𝟏
𝒊 ) − 𝐥𝐨𝐠 (∑ 𝒂𝒊 ))
𝒙
𝒊=𝟏 𝒊=𝟏
𝒏 𝒏 −𝟏 𝒏 𝒏 −𝟏
𝐥𝐢𝐦 𝒈(𝒙) = 𝐥𝐢𝐦 (∑ 𝒂𝒙+𝟏

𝒊 ⋅ 𝐥𝐨𝐠 𝒂𝒊 ) ⋅ (∑ 𝒂𝒙+𝟏
𝒊 ) = (∑ 𝒂𝒊 ⋅ 𝐥𝐨𝐠 𝒂𝒊 ) ⋅ (∑ 𝒂𝒊 ) =
𝒙→𝟎 𝒙→𝟎
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏 𝒊=𝟏
−𝟏
𝒏 𝒏 −𝟏 𝒏 (∑𝒏
𝒊=𝟏 𝒂𝒊 )
𝒂 𝒂
= (∑ 𝐥𝐨𝐠 𝒂𝒊 𝒊 ) ⋅ (∑ 𝒂𝒊 ) = 𝐥𝐨𝐠 (∏ 𝒂𝒊 𝒊 )
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏
Therefore,
𝟏
𝒙𝒏 −𝟏
𝒏 𝒏 −𝟏 𝒏 (∑𝒏
𝒊=𝟏 𝒂𝒊 )
𝟏+𝒙𝒏 𝒂
𝛀 = 𝐥𝐢𝐦 𝐥𝐢𝐦 ((∑ 𝒂𝒊 ) ⋅ (∑ 𝒂𝒊 ) ) = (∏ 𝒂𝒊 𝒊 )
𝒙→𝟎 𝒏→∞
𝒊=𝟏 𝒊=𝟏 𝒊=𝟏
( )
www.ssmrmh.ro
UP.392 Find
𝒏
𝟏 𝟐
𝐥𝐢𝐦 𝒏
⋅ ∑ [(√𝒌 + 𝟏 + √𝒌) ]
𝒏→∞ √(𝒏!)((𝟐𝒏 − 𝟏)‼) 𝒌=𝟏
where [∗] great integer function.

Proposed by D.M. Bătinețu-Giurgiu, Daniel Sitaru-Romania
𝒏
𝟏 𝟐
𝐥𝐢𝐦 ⋅ ∑ [(√𝒌 + 𝟏 + √𝒌) ] =
𝒏→∞ 𝒏
√(𝒏!)((𝟐𝒏 − 𝟏)‼) 𝒌=𝟏
𝒏
𝒏𝟐 𝟏 𝟐
= 𝐥𝐢𝐦 ⋅ 𝟐 ⋅ ∑ [(√𝒌 + 𝟏 + √𝒌) ] ; (𝟏)
𝒏→∞ 𝒏 𝒏
√(𝒏!)((𝟐𝒏 − 𝟏)‼) 𝒌=𝟏
𝒏𝟐 𝒏 𝒏𝟐𝒏 𝑪−𝑫′ 𝑨
𝐥𝐢𝐦 = 𝐥𝐢𝐦 √ =
𝒏→∞ 𝒏 𝒏→∞ 𝒏! ⋅ ((𝟐𝒏 − 𝟏)‼)
√(𝒏!)((𝟐𝒏 − 𝟏)‼)
(𝒏 + 𝟏)𝟐𝒏+𝟐 𝒏! ⋅ ((𝟐𝒏 − 𝟏)‼) 𝒏+𝟏 𝒏 + 𝟏 𝟐𝒏 𝒆𝟐

= 𝐥𝐢𝐦 ⋅ = 𝐥𝐢𝐦 ⋅ ( ) =
𝒏→∞ (𝒏 + 𝟏)! ((𝟐𝒏 + 𝟏)‼) 𝒏𝟐𝒏 𝒏→∞ 𝟐𝒏 + 𝟏 𝒏 𝟐
𝒏 𝟐
𝟏 𝟐 𝑪−𝑺 [(√𝒏 + 𝟐 + √𝒏 + 𝟏) ]
𝐥𝐢𝐦 𝟐 ⋅ ∑ [(√𝒌 + 𝟏 + √𝒌) ] = 𝐥𝐢𝐦
𝒏→∞ 𝒏 𝒏→∞ (𝒏 + 𝟏)𝟐 − 𝒏𝟐
𝒌=𝟏
𝟐
[(√𝒏 + 𝟐 + √𝒏 + 𝟏) ]
= 𝐥𝐢𝐦 ; ( 𝟐)
𝒏→∞ 𝟐𝒏 + 𝟏
We have:
𝟐 𝟐 𝟐
[(√𝒏 + 𝟐 + √𝒏 + 𝟏) ] ≤ (√𝒏 + 𝟐 + √𝒏 + 𝟏) < [(√𝒏 + 𝟐 + √𝒏 + 𝟏) ] + 𝟏, ∀𝒏 ≥ 𝟏
Hence,
𝟐 𝟐 𝟐
[(√𝒏 + 𝟐 + √𝒏 + 𝟏) ] (√𝒏 + 𝟐 + √𝒏 + 𝟏) [(√𝒏 + 𝟐 + √𝒏 + 𝟏) ] 𝟏
≤ < + , ∀𝒏
𝟐𝒏 + 𝟏 𝟐𝒏 + 𝟏 𝟐𝒏 + 𝟏 𝟐𝒏 + 𝟏
≥𝟏
𝟐 𝟐
[(√𝒏 + 𝟐 + √𝒏 + 𝟏) ] [(√𝒏 + 𝟐 + √𝒏 + 𝟏) ] 𝟏
𝐥𝐢𝐦 ≤ 𝟐 < 𝐥𝐢𝐦 + 𝐥𝐢𝐦 ⇒
𝒏→∞ 𝟐𝒏 + 𝟏 𝒏→∞ 𝟐𝒏 + 𝟏 𝒏→∞ 𝟐𝒏 + 𝟏

www.ssmrmh.ro
𝟐
[(√𝒏 + 𝟐 + √𝒏 + 𝟏) ]
𝐥𝐢𝐦 =𝟐
𝒏→∞ 𝟐𝒏 + 𝟏
Therefore,
𝒏
𝒏𝟐 𝟏 𝟐 (𝟏) 𝒆
𝟐
= 𝐥𝐢𝐦 ⋅ 𝟐 ⋅ ∑ [(√𝒌 + 𝟏 + √𝒌) ] = ⋅ 𝟐 = 𝒆𝟐
𝒏→∞ 𝒏 𝒏 𝟐
√(𝒏!)((𝟐𝒏 − 𝟏)‼) 𝒌=𝟏
Solution 2 by Syed Shahabudeen-Kerala-India

𝒏 𝒏
𝟏 𝟐 𝟐 𝟐
𝛀 = 𝐥𝐢𝐦 ∑ [(√𝒌 + 𝟏 + √𝒌) ] = 𝐥𝐢𝐦 𝒏 ∑ [(√𝒌 + 𝟏 + √𝒌) ] =
𝒏→∞ 𝒏√(𝒏!)(𝟐𝒏 − 𝟏)‼ 𝒌=𝟏 𝒏→∞ √𝟐𝒏! 𝒌=𝟏
(𝟐𝒏)!
∵ (𝟐𝒏 − 𝟏)‼ =
𝟐𝒏 𝒏!
𝒏
𝟏 𝟐𝒏𝟐 𝟐
= 𝐥𝐢𝐦 𝒏 ⋅ 𝟐 ∑ [(√𝒌 + 𝟏 + √𝒌) ] = 𝑳 ⋅ 𝑺
𝒏→∞ √𝟐𝒏! 𝒏
𝒌=𝟏
𝟐
𝒏 𝑺𝒕𝒊𝒓𝒍𝒊𝒏𝒈 𝒏𝟐 𝟏 𝒆𝟐 𝒆𝟐
𝑳 = 𝟐 𝐥𝐢𝐦 = 𝟐 𝐥𝐢𝐦 = 𝐥𝐢𝐦 =
𝒏→∞ 𝒏√𝟐𝒏! 𝒏→∞ 𝟏 𝟐𝒏 𝟐 𝟐 𝒏→∞
(𝟒𝒏𝝅)𝟐𝒏
𝟏 𝟐
(𝟒𝒏𝝅)𝟐𝒏 (
𝒆)
𝒏 𝟐
𝟏 𝟐 𝑺𝒕𝒐𝒍𝒛 [(√𝒏 + 𝟏 + √𝒏) ]
𝑺 = 𝐥𝐢𝐦 𝟐 ∑ [(√𝒌 + 𝟏 + √𝒌) ] = 𝐥𝐢𝐦 ≤
𝒏→∞ 𝒏 𝒏→∞ 𝟐𝒏 + 𝟏
𝒌=𝟏
𝟐𝒏 + 𝟏 + 𝟐√𝒏(𝒏 + 𝟏)
= 𝐥𝐢𝐦 =𝟐
𝒏→∞ 𝟐𝒏 + 𝟏
Therefore,
𝒏
𝟏 𝟐 𝒆𝟐
𝛀 = 𝐥𝐢𝐦 ∑(√𝒌 + 𝟏 + √𝒌) = ⋅ 𝟐 = 𝒆𝟐
𝒏→∞ 𝒏√(𝒏!)(𝟐𝒏 − 𝟏)‼ 𝒌=𝟏 𝟐
Solution 3 by Kaushik Mahanta-Assam-India

𝒏 𝒏 𝒏
𝟐
∑[(√𝒌 + 𝟏 + √𝒌) ] = ∑ (𝟐𝒌 + 𝟏 + [√(𝟐𝒌 + 𝟏) 𝟐 − 𝟏]) = ∑ (𝟒𝒌 + 𝟏) = 𝟐𝒏𝟐 + 𝟑𝒏
𝒌=𝟏 𝒌=𝟏 𝒌=𝟏
𝒏
𝟏 𝟐 𝟏
𝛀 = 𝐥𝐢𝐦 ∑(√𝒌 + 𝟏 + √𝒌) = 𝐥𝐢𝐦 (𝟐𝒏𝟐 + 𝟑𝒏) =
𝒏→∞ 𝒏√(𝒏!)(𝟐𝒏 − 𝟏)‼ 𝒌=𝟏 𝒏→∞ 𝒏√(𝒏!)(𝟐𝒏 − 𝟏) ‼

www.ssmrmh.ro
𝟐𝒏𝟐 + 𝟑𝒏 𝟐(𝟐𝒏𝟐 + 𝟑𝒏)
= 𝐥𝐢𝐦 = 𝐥𝐢𝐦 𝒏 =
𝒏→∞ 𝒏 (𝟐𝒏)! 𝒏→∞ √ ( 𝟐𝒏) !
√(𝒏!) ( 𝒏 )
𝟐 (𝒏!)
𝟐𝒏 𝟐𝒏
∵ (𝟐𝒏)! = √𝟐𝝅 ⋅ 𝟐𝒏 ( )
𝒆
𝟐(𝟐𝒏𝟐 + 𝟑𝒏) 𝟐(𝟐𝒏𝟐 + 𝟑𝒏) 𝟐
= 𝐥𝐢𝐦 𝟐𝒏 = 𝐥𝐢𝐦 ⋅ 𝒆 = 𝒆𝟐
𝒏→∞ 𝟏 𝟐𝒏 𝒏 𝒏→∞ (𝟐𝒏)𝟐
(𝟐𝝅 ⋅ 𝟐𝒏)𝟐𝒏 ( ) 𝒆
𝒏
𝟐
𝐋𝐞𝐭: 𝒂𝒌 = [(√𝒌 + √𝒌 + 𝟏) ] , 𝒃𝒏 = ∑ 𝒂𝒌 . 𝐖𝐞 𝐡𝐚𝐯𝐞:
𝒌=𝟏
𝟐𝒌 < 2√𝒌(𝒌 + 𝟏) < 2𝒌 + 𝟏 ⇔ 𝟒𝒌𝟐 < 4𝒌(𝒌 + 𝟏) < 4𝒌𝟐 + 𝟒𝒌 + 𝟏

𝟐 𝟐
𝟒𝒌 + 𝟏 < (√𝒌 + √𝒌 + 𝟏) < 4𝒌 + 𝟐 ⇔ [(√𝒌 + √𝒌 + 𝟏) ] = 𝟒𝒌 + 𝟏
𝒏
⇒ 𝒃𝒏 = ∑ 𝒂𝒌 = 𝟐𝒏(𝒏 + 𝟏) + 𝒏 = 𝒏(𝟐𝒏 + 𝟑)
𝒌=𝟏
𝒃𝒏 𝒏 𝒏𝒏 (𝟐𝒏 + 𝟑)𝒏 𝒏 𝒏𝒏 (𝟐𝒏 + 𝟑)𝒏

𝐋𝐞𝐭: 𝒄𝒏 = =√ = √𝒅𝒏 , 𝐰𝐡𝐞𝐫𝐞 𝒅𝒏 = ; ∀𝒏 ∈ ℕ∗
𝒏
√𝒏! (𝟐𝒏 + 𝟏)‼ 𝒏! (𝟐𝒏 + 𝟏)‼ 𝒏! (𝟐𝒏 + 𝟏)‼
𝒅𝒏+𝟏 (𝒏 + 𝟏)𝒏+𝟏 (𝟐𝒏 + 𝟓)𝒏+𝟏 𝒏! (𝟐𝒏 + 𝟏)‼ 𝟏 𝒏 𝟐 𝒏+𝟏

= ⋅ 𝒏 = (𝟏 + ) (𝟏 + ) =
𝒅𝒏 (𝒏 + 𝟏)! (𝟐𝒏 + 𝟑)‼ 𝒏 (𝟐𝒏 + 𝟑)𝒏 𝒏 𝟐𝒏 + 𝟑
𝟐𝒏+𝟏
𝟐𝒏+𝟑 𝟐𝒏+𝟑
𝒏
𝟏 𝟐 𝟐 𝒅𝒏+𝟏
= (𝟏 + ) [(𝟏 + ) ] ⇒ 𝐥𝐢𝐦 = 𝒆𝟐
𝒏 𝟐𝒏 + 𝟑 𝒏→∞ 𝒅𝒏
We show that
𝒅𝒏+𝟏
𝐥𝐢𝐦 𝒏√𝒅𝒏 = 𝐥𝐢𝐦 = 𝒆𝟐
𝒏→∞ 𝒏→∞ 𝒅𝒏
Given 𝟎 < 𝜀 < 1 there exists a positive integer 𝒎 such that:

𝒅𝒏+𝟏
| − 𝒆𝟐 | < 𝜀, ∀𝑛 > 𝑚
𝒅𝒏
𝒅 𝒅𝒎 𝒅 𝒅𝒎
Let: 𝒌𝟏 = 𝐦𝐢𝐧 {𝒅𝟏 , 𝒅𝟐 , … , 𝒅 } and 𝒌𝟐 = 𝐦𝐚𝐱 {𝒅𝟏 , 𝟐 , … , }
𝟏 𝒎−𝟏 𝒅 𝟏 𝒅 𝒎−𝟏

www.ssmrmh.ro
𝒅
Also, for 𝒏 > 𝑚: 𝒆𝟐 − 𝜺 < 𝒅𝒏+𝟏 > 𝒆𝟐 + 𝜺
𝒏
𝒅 𝒅𝒎 𝒅𝒎+𝟏 𝒅𝒏
For 𝒏 > 𝑚: 𝒌𝒎 𝟐
𝟏 (𝒆 − 𝜺 )
𝒏−𝒎
≤ 𝒅𝟏 ⋅ 𝒅𝟐 ⋅ … ⋅ 𝒅 ⋅ ⋅…⋅ 𝒅 ≤ 𝒌𝒎 𝟐
𝟐 (𝒆 + 𝜺)
𝒏−𝒎
𝟏 𝒎−𝟏 𝒅𝒎 𝒏−𝟏
𝒎 𝒎 𝟏 𝒎 𝒎
⇒ 𝒌𝟏𝒏 (𝒆𝟐 − 𝜺)𝟏− 𝒏 ≤ 𝒏
𝒅𝒏 ≤ 𝒌𝟐𝒏 (𝒆𝟐 + 𝜺)𝟏− 𝒏
𝒆𝟐 − 𝜺 ≤ 𝐥𝐢𝐦 𝒏√𝒅𝒏 ≤ 𝒆𝟐 + 𝜺 ⇒ 𝐥𝐢𝐦 𝒏√𝒅𝒏 = 𝒆𝟐

𝒏→∞ 𝒏→∞
UP.393 Calculate the integral:

𝐭𝐚𝐧−𝟏 𝒙 ∞
𝛀=∫ 𝟒 𝒅𝒙
𝟎 𝒙 +𝟏
It is required to express the integral value with the usual mathematical
𝟑
constants and 𝝍𝟏 ( ) , where 𝝍𝟏 (𝒙) is the trigamma function.
𝟖
Proposed by Vasile Mircea Popa-Romania
Consider the integral with parameter:
∞ ∞
𝐭𝐚𝐧−𝟏 (𝒂𝒙) ′( )
𝒙
( )
𝛀 𝐚 = ∫ 𝒅𝒙 ⇒ 𝛀 𝐚 = ∫ 𝒅𝒙
𝟎 𝒙𝟒 + 𝟏 𝟒 𝟐 𝟐
𝟎 (𝒙 + 𝟏)(𝟏 + 𝒂 𝒙 )
We make notations: 𝒂𝟐 = 𝒃 and we have:
∞ ∞
𝒙 𝒖=𝒙𝟐 ;𝒅𝒖=𝟐𝒙𝒅𝒙 𝟏 𝟏
𝛀𝟏 (𝒂) = ∫ 𝟒 𝟐
𝒅𝒙 = ∫ 𝟐
𝒅𝒖
𝟎 (𝒙 + 𝟏)(𝟏 + 𝒃𝒙 ) 𝟐 𝟎 (𝒖 + 𝟏)(𝟏 + 𝒃𝒖)
𝟏 𝑨 𝑩𝒖+𝑪
We can write: (𝒖𝟐 +𝟏)(𝟏+𝒃𝒖) = 𝟏+𝒃𝒖 + 𝒖𝟐 +𝟏 , where 𝑨, 𝑩, 𝑪 have the following values:
𝒃𝟐 𝒃 𝟏
𝑨= 𝟐
,𝑩 = − 𝟐 ,𝑪 = 𝟐
𝒃 +𝟏 𝒃 +𝟏 𝒃 +𝟏
We have:
𝑨 𝑨
∫ 𝒅𝒖 = 𝐥𝐨𝐠(𝟏 + 𝒃𝒖) + 𝑲
𝟏 + 𝒃𝒖 𝒃
𝑩𝒖 + 𝑪 𝑩
∫ 𝟐 𝒅𝒖 = 𝐥𝐨𝐠(𝒖𝟐 + 𝟏) + 𝑪 𝐭𝐚𝐧−𝟏 𝒖 + 𝑲
𝒖 +𝟏 𝟐
Hence,
∞
𝟏 𝑨 𝑩
∫ 𝟐 + 𝟏)(𝟏 + 𝒃𝒖)
𝒅𝒖 = 𝐥𝐨𝐠 ( 𝟏 + 𝒃𝒖 ) + 𝐥𝐨𝐠(𝒖𝟐 + 𝟏) + 𝑪 𝐭𝐚𝐧−𝟏 𝒖 + 𝑲
𝟎 ( 𝒖 𝒃 𝟐
We make notation:
𝟏
𝑷(𝒖) = ∫ 𝟐 𝒅𝒖 = 𝑸(𝒖) + 𝑹(𝒖), 𝐰𝐡𝐞𝐫𝐞: 𝑸(𝒖) = 𝑪 𝐭𝐚𝐧−𝟏 𝒖
(𝒖 + 𝟏)(𝟏 + 𝒃𝒖)
𝑨 𝑩
𝑹(𝒖) = 𝐥𝐨𝐠(𝟏 + 𝒃𝒖) + 𝐥𝐨𝐠(𝒖𝟐 + 𝟏)
𝒃 𝟐
We calculate:

www.ssmrmh.ro
𝝅 𝟏
𝚫𝟏 = 𝐥𝐢𝐦 𝑸(𝒖) − 𝑸(𝟎) , 𝐰𝐞 𝐨𝐛𝐭𝐚𝐢𝐧: 𝚫𝟏 = ⇒𝑪= 𝟐
𝒖→∞ 𝟐(𝒃𝟐 + 𝟏) 𝒃 +𝟏
We calculate:
𝚫𝟐 = 𝐥𝐢𝐦 𝑹(𝒖) − 𝑹(𝟎)
𝒖→∞
𝒃 𝒃𝒖+𝟏 𝑨 𝒃 𝑩 𝒃
We can write: 𝑹(𝒖) = 𝒃𝟐 +𝟏 𝐥𝐨𝐠 ( ) because 𝒃 = 𝒃𝟐 +𝟏 and 𝟐 = − 𝟐(𝒃𝟐 +𝟏).
√𝒖𝟐 +𝟏
𝒃
We obtain: 𝚫𝟐 = 𝒃𝟐 +𝟏 𝐥𝐨𝐠 𝒃.
Therefore, using Newton-Leibniz formula, we can write:
∞
𝟏 𝝅 + 𝟐𝒃 𝐥𝐨𝐠 𝒃
∫ 𝟐
𝒅𝒖 = 𝚫𝟏 + 𝚫𝟐 = , 𝐨𝐫:
𝟎 (𝒖 + 𝟏)(𝟏 + 𝒃𝒖) 𝟐 + 𝟐𝒃𝟐
∞
𝟏 𝝅 + 𝟐𝒂𝟐 𝐥𝐨𝐠 𝒃
∫ 𝟐 𝟐
𝒅𝒖 =
𝟎 (𝒖 + 𝟏)(𝟏 + 𝒂 𝒖) 𝟐 + 𝟐𝒂𝟒
𝟐
𝟏 𝝅 + 𝟐𝒂 𝐥𝐨𝐠 𝒃
𝛀𝟏 (𝒂) = ⋅ , 𝐛𝐞𝐜𝐚𝐮𝐬𝐞 𝛀(𝟎) = 𝟎 𝐰𝐞 𝐡𝐚𝐯𝐞:
𝟐 𝟐(𝟏 + 𝒂𝟒 )
𝟏
𝛀 = 𝛀(𝟏) = ∫ 𝛀𝟏 (𝒂) 𝒅𝒂
𝟎
𝝅 𝟏 𝟏 𝟏 𝟐
𝒂 𝐥𝐨𝐠 𝒂
𝛀= ∫ 𝒅𝒂 + ∫ 𝒅𝒂 ; (𝟏)
𝟒 𝟎 𝟏 + 𝒂𝟒 𝟎 𝟏+𝒂
𝟒
The first integral in the above relation are integral of rational function and are
calculated relatively easily with the Newton-Leibniz formula. The integral have the
following value:
𝟏
𝟏 √𝟐
∫ 𝟒
𝒅𝒂 = (𝝅 + 𝟐 𝐥𝐨𝐠(𝟏 + √𝟐)); (𝟐)
𝟎 𝟏+𝒂 𝟖
𝟏 𝒙𝟐 𝐥𝐨𝐠 𝒙
We calculate now the second integral: ∫𝟎 𝒅𝒙
𝟏+𝒙𝟒
𝒙𝟐
We make the notation: 𝒇(𝒙) = 𝟏+𝒙𝟒 and develop the function in power series.
We have for 𝒙 ∈ (𝟎, 𝟏): 𝒇(𝒙) = 𝒙𝟐 − 𝒙𝟔 + 𝒙𝟏𝟎 − 𝒙𝟏𝟒 + 𝒙𝟏𝟖 − 𝒙𝟐𝟐 + 𝒙𝟐𝟔 − 𝒙𝟑𝟎 +
𝒙𝟑𝟒 − 𝒙𝟑𝟖 …. We obtain:
𝟏
𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
∫ 𝒇(𝒙) 𝐥𝐨𝐠 𝒙 𝒅𝒙 = − 𝟐 + 𝟐 − 𝟐 + 𝟐 − 𝟐 + 𝟐 − 𝟐 + 𝟐 − 𝟐 + 𝟐 − ⋯
𝟎 𝟑 𝟕 𝟏𝟏 𝟏𝟓 𝟏𝟗 𝟐𝟑 𝟐𝟕 𝟑𝟏 𝟑𝟓 𝟑𝟗
𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
= −( 𝟐 + 𝟐 + 𝟐 + 𝟐 + 𝟐 +⋯) + ( 𝟐 + 𝟐 + 𝟐 + 𝟐 + 𝟐 + ⋯)
𝟑 𝟏𝟏 𝟏𝟗 𝟐𝟕 𝟑𝟓 𝟕 𝟏𝟓 𝟐𝟑 𝟑𝟏 𝟑𝟗
Now we will use the trigamma function, which is defined by the relationship:
∞
𝟏
𝝍𝟏 ( 𝒙 ) = ∑
(𝒙 + 𝒏)𝟐
𝒏=𝟎
∞ ∞
𝟑 𝟏 𝟕 𝟏
𝐖𝐞 𝐜𝐚𝐧 𝐰𝐫𝐢𝐭𝐞: 𝛙𝟏 ( ) = ∑ 𝟐 ; 𝛙𝟐 ( ) = ∑ 𝟐
𝟖 𝟑 𝟖 𝟕
𝐧=𝟎 ( + 𝐧) 𝐧=𝟎 ( + 𝐧)
𝟖 𝟖

www.ssmrmh.ro
𝟏
𝟏 𝟑 𝟕
𝐒𝐨, ∫ 𝒇(𝒙) 𝐥𝐨𝐠 𝒙 𝒅𝒙 = (−𝝍𝟏 ( ) + 𝝍𝟏 ( ))
𝟎 𝟔𝟒 𝟖 𝟖
But we have the following equality (the duplication formula):
𝟏 𝟑
𝟒𝝍𝟏 (𝟐𝒙) = 𝝍𝟏 (𝒙) + 𝝍𝟏 (𝒙 + ) , 𝐰𝐡𝐢𝐜𝐡 𝐟𝐨𝐫 𝒙 = 𝐰𝐞 𝐠𝐢𝐯𝐞𝐬 𝐮𝐬:
𝟐 𝟖
𝟑 𝟑 𝟕
𝟒𝝍𝟏 ( ) = 𝝍𝟏 ( ) + 𝝍𝟏 ( )
𝟒 𝟖 𝟖
The following special value is known:
𝟑
𝝍𝟏 ( ) = 𝝅𝟐 − 𝟖𝑮, 𝐰𝐡𝐞𝐫𝐞 𝐆 − 𝒊𝒔 𝑪𝒂𝒕𝒂𝒍𝒂𝒏′ 𝒔 𝐜𝐨𝐧𝐬𝐭𝐚𝐧𝐭.
𝟒
Therefore, we can write:
𝟏 𝟐
𝒙 𝐥𝐨𝐠 𝒙 𝟏 𝟑
∫ 𝟒 𝒅𝒙 = (−𝟐𝝍𝟏 ( ) + 𝟒𝝅𝟐 − 𝟑𝟐𝑮) ; (𝟑)
𝟎 𝒙 +𝟏 𝟔𝟒 𝟖
Replacing the values of the integrals (2) and (3) in the relation (1), we obtain the
value of the integral from the problem statement. This is:
√𝟐 𝝅√𝟐 𝟏 𝟑 𝟏
𝛀= (𝟏 + √𝟐)𝝅𝟐 + 𝐥𝐨𝐠(𝟏 + √𝟐) − 𝝍𝟏 ( ) − 𝑮
𝟑𝟐 𝟏𝟔 𝟑𝟐 𝟖 𝟐
Solution 2 by Rana Ranino-Setif-Algerie

∞ ∞ ∞
𝐭𝐚𝐧−𝟏 𝒙 𝒙
𝛀=∫ 𝟒
𝒅𝒙 = ∫ ∫ 𝟒 𝟐 𝟐
𝒅𝒙𝒅𝒚 =
𝟎 𝒙 +𝟏 𝟎 𝟎 (𝟏 + 𝒙 )(𝟏 + 𝒙 𝒚 )
𝟏 𝟏 𝟏 ∞
𝒙 𝒙𝟑 𝒚𝟐 𝒙𝒚𝟒
= ∫ ∫ ( − + ) 𝒅𝒙𝒅𝒚 =
𝟐 𝟎 𝟏 + 𝒚𝟒 𝟎 𝟏 + 𝒙 𝟒 𝟏 + 𝒙 𝟒 𝟏 + 𝒙 𝟐 𝒚 𝟐
∞
𝟏 𝟏 𝟏 −𝟏 𝟐 𝟐 𝟐 𝟐
𝟏 𝟐 𝟒
= ∫ ( ) ( ) ( )
[𝐭𝐚𝐧 𝒙 + 𝒚 𝐥𝐨𝐠 𝟏 + 𝒙 𝒚 − 𝒚 𝐥𝐨𝐠 𝟏 + 𝒙 ] 𝒅𝒚 =
𝟐 𝟎 𝟏 + 𝒚𝟒 𝟐 𝟎
𝟏 𝟏 𝟏 𝝅
= ∫ 𝟒
( + 𝟐𝒚𝟐 𝐥𝐨𝐠 𝒚) 𝒅𝒚
𝟐 𝟎 𝟏+𝒚 𝟐
𝝅 𝟏 𝒅𝒚 𝟏 𝟐
𝒚 𝐥𝐨𝐠 𝒚
𝛀= ∫ + ∫ 𝒅𝒚
𝟒 ⏟𝟎 𝟏 + 𝒚𝟒 ⏟𝟎 𝟏 + 𝒚𝟒
𝑨 𝑩
𝟏 𝟏 𝒚𝟐 + 𝟏 𝟏 𝟏 𝒚𝟐 − 𝟏
𝑨= ∫ 𝒅𝒚 − ∫ 𝒅𝒚
𝟐 𝟎 𝟏 + 𝒚𝟒 𝟐 𝟎 𝟏 + 𝒚𝟒
𝟏 𝟏 𝟏 𝟏
𝟏 𝟏 𝟐 𝟏+ 𝟏 𝟏 𝟏− 𝟐 𝒖=𝒚− ; 𝒗=𝒚+
𝒚 𝒚
𝒚 𝒚
= ∫ 𝒅𝒚 − ∫ 𝒅𝒚 =
𝟐 𝟎 𝒚𝟐 + 𝟏 𝟐 𝟎 𝒚𝟐 + 𝟏
𝒚𝟐 𝒚𝟐

www.ssmrmh.ro
𝟎
𝟏 𝒅𝒖 𝟏 ∞ 𝒅𝒗 𝝅 𝐥𝐨𝐠(𝟑 − 𝟐√𝟐)
= ∫ 𝟐
+ ∫ 𝟐 = −
𝟐 −∞ 𝒖 + 𝟐 𝟐 𝟐 𝒗 − 𝟏 𝟒√𝟐 𝟒√𝟐
𝟑 𝟕
−𝟏
𝟏( 𝟐
𝒚 − 𝒚𝟔 ) 𝐥𝐨𝐠 𝒚 𝒚→𝒚𝟖 𝟏 𝟏 (𝒚𝟖 − 𝒚𝟖−𝟏 ) 𝐥𝐨𝐠 𝒚
𝑩=∫ 𝒅𝒚 = ∫ 𝒅𝒚 =
𝟎 𝟏 − 𝒚𝟖 𝟔𝟒 𝟎 𝟏−𝒚
𝟏 𝟕 𝟑
= {𝝍(𝟏) ( ) − 𝝍(𝟏) ( )}
𝟔𝟒 𝟖 𝟖
Therefore,
∞
𝐭𝐚𝐧−𝟏 𝒙 𝝅𝟐 𝝅 𝐥𝐨𝐠(𝟑 − 𝟐√𝟐) 𝟏 𝟕 𝟑
𝛀=∫ 𝟒
𝒅𝒙 = − + {𝝍(𝟏) ( ) − 𝝍(𝟏) ( )}
𝟎 𝒙 +𝟏 𝟏𝟔√𝟐 𝟏𝟔√𝟐 𝟔𝟒 𝟖 𝟖
UP.394
𝒂 𝒙
𝛀(𝒂) = ∫ 𝒅𝒙, 𝒂 > 2
𝟏 𝐥𝐨𝐠(𝟏 + 𝒙𝟐 )
Prove that:
(𝒂 + 𝒃 + 𝒄)𝟐 𝟏 𝒂𝟐 𝒃𝟐 𝒄𝟐 𝟐
𝒂𝟐 𝒃𝟐 𝒄𝟐
+ 𝐥𝐨𝐠 ( ) ≤ ∑ 𝛀(𝒂) ≤ (𝒂 + 𝒃 + 𝒄) + 𝐥𝐨𝐠 ( )
𝟑√𝟐 √𝟐 𝒆𝟑 𝒆√𝒆
𝒄𝒚𝒄

𝒙 𝒙
− 𝟏
𝒙−𝒚 𝒙+𝒚 𝒚 𝒚+𝟏 𝒙
∵ ≤ ; ∀𝒙, 𝒚 > 0, 𝑙𝑒𝑡 𝑥 > 𝑦 ⇒ 𝒙 ≤ ; 𝒕 = >1⇒
𝐥𝐨𝐠 𝒙 − 𝐥𝐨𝐠 𝒚 𝟐 𝐥𝐨𝐠 (𝒚) 𝟐 𝒚
𝒕−𝟏 𝒕+𝟏 𝒕 𝒕
< , ∀𝒕 > 1; (𝒕 → 𝒕 + 𝟏) ⟺ < + 𝟏;
𝐥𝐨𝐠 𝒕 𝟐 𝐥𝐨𝐠(𝟏 + 𝒕) 𝟐
𝟐
𝒙𝟐 𝟏
(𝒕 = 𝒙 ) ⇒ 𝟐
< 𝒙𝟐 + 𝟏 ⟺
𝐥𝐨𝐠(𝟏 + 𝒙 ) 𝟐
𝒙 𝟏 𝟏
𝟐
< 𝒙 + ; ∀𝒙 > 0; (𝟏)
𝐥𝐨𝐠(𝟏 + 𝒙 ) 𝟐 𝒙
𝒕 𝒕
∵ 𝐥𝐨𝐠(𝟏 + 𝒕) ≤ ; ∀𝒕 > 0 ⟺ √𝟏 + 𝒕 ≤ ; ∀𝒕 > 0 ⇒
√𝟏 + 𝒕 𝐥𝐨𝐠(𝟏 + 𝒕)
𝒙𝟐 𝟏+𝒙 𝟏 𝟏 𝒙
√𝟏 + 𝒙𝟐 ≤ 𝐚𝐧𝐝 √𝟏 + 𝒙 𝟐 ≥ ⇒ (𝒙 + ) ≤ ; ∀𝒙
𝐥𝐨𝐠(𝟏 + 𝒙𝟐 ) √𝟐 √𝟐 𝒙 𝐥𝐨𝐠(𝟏 + 𝒙𝟐 )
> 0; (𝟐)
www.ssmrmh.ro
From (1),(2) we get:
𝟏 𝟏 𝒙 𝒙 𝟏
(𝒙 + ) < < +
√𝟐 𝒙 𝐥𝐨𝐠(𝟏 + 𝒙𝟐 ) 𝟐 𝒙
𝒂 𝒂
𝟏 𝟏 𝒙 𝟏
∫ (𝒙 + ) 𝒅𝒙 ≤ 𝛀(𝒂) ≤ ∫ ( + ) 𝒅𝒙 ⟺
√𝟐 𝟏 𝒙 𝟏 𝟐 𝒙
𝒂 𝒂
𝒙𝟐
𝟏 𝒙𝟐
( + 𝐥𝐨𝐠 𝒙)| ≤ 𝛀(𝒂) ≤ ( + 𝐥𝐨𝐠 𝒙)| ⟺
√𝟐 𝟐 𝟏
𝟒 𝟏
𝒂𝟐
𝟏 𝟏 𝒂𝟐 𝟏
( + 𝐥𝐨𝐠 𝒂 − ) ≤ 𝛀(𝒂) ≤ + 𝐥𝐨𝐠 𝒂 −
√𝟐 𝟐 𝟐 𝟒 𝟒
𝟏 𝟐 𝟐
𝒂𝟐 𝟏
(𝒂 + 𝐥𝐨𝐠 𝒂 − 𝟏) ≤ 𝛀(𝒂) ≤ + 𝐥𝐨𝐠 𝒂𝟐 −
√𝟐 𝟐 𝟐
𝟏 𝒂𝟐 𝒂𝟐 𝒂𝟐
(𝒂𝟐 + 𝐥𝐨𝐠 ( )) ≤ 𝛀(𝒂) ≤ + 𝐥𝐨𝐠 ( )
√𝟐 𝒆 𝟐 √𝒆
Adding, we get:
𝟏 𝒂𝟐 𝒃𝟐 𝒄𝟐 𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 𝒂𝟐 𝒃𝟐 𝒄𝟐
(𝒂𝟐 + 𝒃𝟐 + 𝒄𝟐 + 𝐥𝐨𝐠 ( )) ≤ ∑ 𝛀 ( 𝒂) ≤ + 𝐥𝐨𝐠 ( )⟺
√𝟐 𝒆𝟑 𝟐 𝒆 √𝒆
𝒄𝒚𝒄
(𝒂 + 𝒃 + 𝒄)𝟐 𝟏 𝒂𝟐 𝒃𝟐 𝒄𝟐 𝟐
𝒂𝟐 𝒃𝟐 𝒄𝟐
+ 𝐥𝐨𝐠 ( ) ≤ ∑ 𝛀 ( 𝒂 ) ≤ ( 𝒂 + 𝒃 + 𝒄) + 𝐥𝐨𝐠 ( )
𝟑√𝟐 √𝟐 𝒆𝟑 𝒆√𝒆
𝒄𝒚𝒄
UP.395 For 𝒂 > 0, 𝑏 > 1 find:

𝒏𝟐
𝟏
𝒏 (𝒂 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( )
𝒏𝟐 − 𝒌𝒏 + 𝒌𝟐
𝛀 = 𝐥𝐢𝐦 ∑
𝒏→∞ 𝒏 𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌
𝒌=𝟏

−𝟏
𝒌 −𝟏
𝒏−𝒌 −𝟏
𝒏𝟐
∵ 𝐭𝐚𝐧 ( ) + 𝐭𝐚𝐧 ( ) = 𝐭𝐚𝐧 ( 𝟐 )
𝒏 𝒏 𝒏 − 𝒌𝒏 + 𝒌𝟐
Let us denote:
𝒏𝟐
(𝒂 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( ) (𝒂 + 𝒃𝒌 ) (𝐭𝐚𝐧−𝟏 (𝒌) + 𝐭𝐚𝐧−𝟏 (𝒏 − 𝒌))
𝒏𝟐
− 𝒌𝒏 + 𝒌𝟐 𝒏 𝒏
𝒂𝒌 = =
𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌 𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌

www.ssmrmh.ro
Hence,
𝒌 𝒏−𝒌
(𝒂 + 𝒃𝒌 ) (𝐭𝐚𝐧−𝟏 ( ) + 𝐭𝐚𝐧−𝟏 (
𝒂𝒌 + 𝒂𝒏−𝒌 = 𝒏 𝒏 ))
𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌
𝒌 𝒏−𝒌
(𝒂 + 𝒃𝒏−𝒌 ) (𝐭𝐚𝐧−𝟏 ( ) + 𝐭𝐚𝐧−𝟏 (
+ 𝒏 𝒏 )) = 𝐭𝐚𝐧−𝟏 (𝒌) + 𝐭𝐚𝐧−𝟏 (𝒏 − 𝒌)
𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌 𝒏 𝒏
and then,
𝒏−𝟏 𝒏 𝒏−𝟏 𝒏−𝟏
−𝟏
𝒌 𝒏−𝒌 𝒌
𝟐 ∑ 𝒂𝒌 = ∑(𝒂𝒌 + 𝒂𝒏−𝒌 ) = ∑ (𝐭𝐚𝐧 ( ) + 𝐭𝐚𝐧−𝟏 ( )) = 𝟐 ∑ 𝐭𝐚𝐧−𝟏 ( )
𝒏 𝒏 𝒏
𝒌=𝟏 𝒌=𝟏 𝒌=𝟏 𝒌=𝟏
Therefore,
𝒏 𝒏−𝟏 𝒏−𝟏
𝒌 (𝒂 + 𝒃𝒏 ) 𝐭𝐚𝐧−𝟏 (𝟏)
∑ 𝒂𝒌 = ∑ 𝒂𝒌 + 𝒂𝒏 = ∑ 𝐭𝐚𝐧−𝟏 ( ) + =
𝒏 𝒃𝒏 + 𝟐𝒂 + 𝟏
𝒌=𝟏 𝒌=𝟏 𝒌=𝟏
𝒏−𝟏
𝒌 𝝅 𝒂 + 𝒃𝒏
= ∑ 𝐭𝐚𝐧−𝟏 ( ) + ⋅
𝒏 𝟒 𝟏 + 𝟐𝒂 + 𝒃𝒏
𝒌=𝟏
So, we have:
𝟐
𝒏
𝒏 (𝒂 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( ) 𝒏−𝟏 𝒌 𝝅 𝒂 + 𝒃𝒏
𝒏𝟐 − 𝒌𝒏 + 𝒌𝟐 −𝟏
∑ = ∑ 𝐭𝐚𝐧 ( ) + ⋅
𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌 𝒏 𝟒 𝟏 + 𝟐𝒂 + 𝒃𝒏
𝒌=𝟏 𝒌=𝟏
and
𝒏𝟐
𝟏
𝒏 (𝒂 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( ) 𝟏 𝒏−𝟏 𝒌 𝝅 𝒂 + 𝒃𝒏
𝒏𝟐 − 𝒌𝒏 + 𝒌𝟐 −𝟏
∑ = ∑ 𝐭𝐚𝐧 ( ) + ⋅ =
𝒏 𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌 𝒏 𝒏 𝟒𝒏 𝟏 + 𝟐𝒂 + 𝒃𝒏
𝒌=𝟏 𝒌=𝟏
𝒏
𝟏 𝒌 𝝅 𝝅 𝒂 + 𝒃𝒏
= ∑ 𝐭𝐚𝐧−𝟏 ( ) − + ⋅ =
𝒏 𝒏 𝟒𝒏 𝟒𝒏 𝟏 + 𝟐𝒂 + 𝒃𝒏
𝒌=𝟏
𝒏 𝒏
𝟏 𝒌 𝝅 𝒂 + 𝒃𝒏 𝟏 𝒌 𝝅 𝒂+𝟏
= ∑ 𝐭𝐚𝐧−𝟏 ( ) + ( − 𝟏) = ∑ 𝐭𝐚𝐧 −𝟏
( ) − ⋅
𝒏 𝒏 𝟒𝒏 𝟏 + 𝟐𝒂 + 𝒃𝒏 𝒏 𝒏 𝟒𝒏 𝟏 + 𝟐𝒂 + 𝒃𝒏
𝒌=𝟏 𝒌=𝟏
Therefore,
𝒏𝟐
𝟏
𝒏 (𝒂 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( ) 𝟏
𝒏
𝒌 𝝅 𝒂+𝟏
𝒏𝟐 − 𝒌𝒏 + 𝒌𝟐 −𝟏
𝐥𝐢𝐦 ∑ = 𝐥𝐢𝐦 ( ∑ 𝐭𝐚𝐧 ( ) − ⋅ )
𝒏→∞ 𝒏 𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌 𝒏→∞ 𝒏 𝒏 𝟒𝒏 𝟏 + 𝟐𝒂 + 𝒃𝒏
𝒌=𝟏 𝒌=𝟏
𝟏
𝝅
= ∫ 𝐭𝐚𝐧−𝟏 𝒙 𝒅𝒙 = − 𝐥𝐨𝐠 √𝟐
𝟎 𝟒

www.ssmrmh.ro
𝒏𝟐
𝟏
𝒏 (𝒂 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( )
𝒏𝟐 − 𝒌𝒏 + 𝒌𝟐
𝑰(𝒏) = ∑ =
𝒏 𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌
𝒌=𝟏
𝒏𝟐
𝟏
𝒏−𝟏 (𝒂 + 𝒃𝒏−𝒌 𝐭𝐚𝐧−𝟏 ( )
𝒏𝟐 − 𝒏𝒌 + 𝒌𝟐
= ∑ =
𝒏 𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌
𝒌=𝟎
𝒏𝟐
𝟏
𝒏 (𝒂 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( ) 𝟏 𝝅 𝒂 + 𝒃𝒏 𝟏 𝝅 𝒂+𝟏
𝒏𝟐 − 𝒌𝒏 + 𝒌𝟐
= ∑ + ⋅ ⋅ + ⋅ ⋅
𝒏 𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌 𝒏 𝟒 𝟏 + 𝟐𝒂 + 𝒃𝒏 𝒏 𝟒 𝟏 + 𝟐𝒂 + 𝒃𝒏
𝒌=𝟏
𝒏𝟐
𝟏
𝒏 (𝟐𝒂 + 𝒃𝒏−𝒌 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( ) 𝝅 𝒂 + 𝒃𝒏
𝒏𝟐 − 𝒏𝒌 + 𝒌𝟐
⇒ 𝟐𝑰(𝒏) = ∑ + ⋅ +
𝒏 𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌 𝟒 𝟏 + 𝟐𝒂 + 𝒃𝒏
𝒌=𝟏
𝒏
𝒂+𝟏 𝝅 𝟏 −𝟏
𝒏𝟐 𝝅
+ 𝒏
⋅ = ∑ 𝐭𝐚𝐧 ( 𝟐 𝟐
)+
𝟏 + 𝟐𝒂 + 𝒃 𝟒 𝒏 𝒏 − 𝒏𝒌 + 𝒌 𝟐𝒏
𝒌=𝟏
𝒏
𝟏 −𝟏
𝒏𝟐 𝝅
𝑰(𝒏) = ∑ 𝐭𝐚𝐧 ( 𝟐 ) +
𝟐𝒏 𝒏 − 𝒏𝒌 + 𝒌𝟐 𝟖𝒏
𝒌=𝟏
𝒏 𝟏
𝟏 𝒏𝟐 𝝅 𝟏
𝐥𝐢𝐦 ∑ 𝐭𝐚𝐧−𝟏 ( 𝟐 𝟐
) + = ∫ 𝐭𝐚𝐧−𝟏 ( ) 𝒅𝒙 =
𝒏→∞ 𝟐𝒏 𝒏 − 𝒏𝒌 + 𝒌 𝟖𝒏 𝟎 𝟏 − 𝒙 + 𝒙𝟐
𝒌=𝟏
𝟏
= ∫ (𝐭𝐚𝐧−𝟏 𝒙 + 𝐭𝐚𝐧−𝟏 (𝟏 − 𝒙))𝒅𝒙 =
𝟎
𝟏 𝟏
−𝟏
𝟏 𝟐
𝟏 𝟐 −𝟏
= [𝒙 𝐭𝐚𝐧 𝒙 − 𝐥𝐨𝐠(𝒙 + 𝟏)] + [ 𝐥𝐨𝐠(𝒙 − 𝟐𝒙 + 𝟐) + (𝒙 − 𝟏) 𝐭𝐚𝐧 (𝟏 − 𝒙)] =
𝟐 𝟎 𝟐 𝟎
𝟏
= (𝝅 − 𝟐 𝐥𝐨𝐠 𝟐)
𝟐
Therefore,
𝒏𝟐
𝟏
𝒏 (𝒂 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( ) 𝟏
𝒏𝟐 − 𝒌𝒏 + 𝒌𝟐
𝐥𝐢𝐦 ∑ = (𝝅 − 𝟐 𝐥𝐨𝐠 𝟐)
𝒏→∞ 𝒏 𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌 𝟒
𝒌=𝟏

For 𝟎 ≤ 𝒌 ≤ 𝒏 let

www.ssmrmh.ro
𝒏𝟐
(𝒂 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( )
𝒏𝟐 − 𝒌𝒏 + 𝒌𝟐
𝒂𝒌 =
𝒃𝒌 + 𝟐𝒂 + 𝒃𝒏−𝒌
𝒏𝟐
(𝒂 + 𝒃𝒏−𝒌 ) 𝐭𝐚𝐧−𝟏 ( )
𝒏𝟐
− (𝒏 − 𝒌 )𝒏 + (𝒏 − 𝒌 )𝟐
⇒ 𝒂𝒏−𝒌 =
𝒃𝒏−𝒌 + 𝟐𝒂 + 𝒃𝒏
𝒌 𝒌
𝟏 (𝟏 − 𝒏) + 𝒏
𝒂𝒌 + 𝒂𝒏−𝒌 = 𝐭𝐚𝐧−𝟏 ( ) = 𝐭𝐚𝐧−𝟏 ( )=
𝒌 𝒌 𝒌 𝒌
𝟏 − (𝟏 − ) 𝟏 − (𝟏 − )
𝒏 𝒏 𝒏 𝒏
𝒌 𝒌 𝒏−𝒌 𝒌
= 𝐭𝐚𝐧−𝟏 (𝟏 − ) + 𝐭𝐚𝐧−𝟏 ( ) = 𝐭𝐚𝐧−𝟏 ( ) + 𝐭𝐚𝐧−𝟏 ( )
𝒏 𝒏 𝒏 𝒏
𝒏
𝐋𝐞𝐭: 𝑺 = ∑ 𝒂𝒌 ⇒ 𝑺 + 𝒂𝟎 = 𝒂𝟎 + 𝒂𝟏 + ⋯ + 𝒂𝒏 = 𝒂𝒏 + ⋯ + 𝒂𝟏 + 𝒂𝟎
𝒌=𝟏
𝒏 𝒏
𝒏−𝒌 𝒌
𝟐𝑺 + 𝟐𝒂𝟎 = ∑ (𝒂𝒌 + 𝒂𝒏−𝒌 ) = ∑ (𝐭𝐚𝐧−𝟏 ( ) + 𝐭𝐚𝐧−𝟏 ( )) =
𝒏 𝒏
𝒌=𝟏 𝒌=𝟎
𝒏 𝒏
𝒌 𝒌
= 𝟐 ∑ 𝐭𝐚𝐧−𝟏 ( ) ⇒ 𝑺 + 𝒂𝟎 = ∑ 𝐭𝐚𝐧−𝟏 ( )
𝒏 𝒏
𝒌=𝟎 𝒌=𝟎
Therefore,
𝒏𝟐
𝟏
𝒏 (𝒂 + 𝒃𝒌 ) 𝐭𝐚𝐧−𝟏 ( ) 𝟏
𝝅
𝒏𝟐
− 𝒌𝒏 + 𝒌𝟐
𝐥𝐢𝐦 ∑ 𝒌 𝒏−𝒌
= ∫ 𝐭𝐚𝐧−𝟏 𝒙 𝒅𝒙 = − 𝐥𝐨𝐠 √𝟐
𝒏→∞ 𝒏 𝒃 + 𝟐𝒂 + 𝒃 𝟎 𝟒
𝒌=𝟏
UP.396. Find:
𝟏
𝒂 𝐬𝐢𝐧𝟐 𝒙 ⋅ 𝐭𝐚𝐧 ( )
𝛀(𝒂) = ∫ 𝒙𝟐 𝒅𝒙
𝟏 𝟏 𝟏
𝒂 (𝟏 + 𝒙𝟐 ) (𝐬𝐢𝐧𝟐 ( ) 𝐭𝐚𝐧 ( 𝟐 ) + 𝐬𝐢𝐧𝟐 𝒙 𝐭𝐚𝐧(𝒙𝟐 ))
𝒙 𝒙
𝟏 𝟏 𝟏 𝟏 𝟏
𝒕= ⇒ 𝒙 = , 𝒅𝒙 = − 𝟐 𝒅𝒕, 𝒙 = ⇒ 𝒕 = 𝒂; 𝒙 = 𝒂 ⇒ 𝒕 =
𝒙 𝒕 𝒕 𝒂 𝒂
Hence,

www.ssmrmh.ro
𝟏 𝟏 𝟏
𝒂 𝐬𝐢𝐧𝟐 ( 𝒕 ) 𝐭𝐚𝐧 𝒕𝟐 ⋅ (− 𝟐 ) 𝒅𝒕
𝛀(𝒂) = ∫ 𝒕 =
𝒂 (𝟏 +
𝟏 𝟐 𝒕 𝐭𝐚𝐧𝟐 ( 𝟏 ) + 𝐬𝐢𝐧𝟐 (𝟏) 𝐭𝐚𝐧 𝒕𝟐 )
) (𝐬𝐢𝐧
𝒕𝟐 𝒕𝟐 𝒕
𝟏
𝒂 𝐬𝐢𝐧𝟐 ( 𝒕 ) 𝐭𝐚𝐧 𝒕𝟐
=∫ 𝒅𝒕
𝟏 𝟏 𝟏
𝒂 (𝟏 + 𝒕𝟐 ) (𝐬𝐢𝐧𝟐 𝒕 𝐭𝐚𝐧𝟐 ( 𝟐 ) + 𝐬𝐢𝐧𝟐 ( ) 𝐭𝐚𝐧 𝒕𝟐 )
𝒕 𝒕
Therefore,
𝟏 𝟐 𝟏
𝒂 𝐬𝐢𝐧𝟐 𝒙 𝐭𝐚𝐧 (
𝟐 ) + 𝐬𝐢𝐧 (𝒙 ) 𝐭𝐚𝐧 𝒙
𝟐 𝒂
𝒅𝒙
𝟐𝛀(𝒂) = ∫ 𝒙 𝒅𝒙 = ∫
𝟏 𝟐 𝟐 𝟏 𝟐 𝟏 𝟐 𝟏 𝟏 + 𝒙𝟐
𝒂 (𝟏 + 𝒙 ) (𝐬𝐢𝐧 𝒙 𝐭𝐚𝐧 ( 𝟐 ) + 𝐬𝐢𝐧 (𝒙 ) 𝐭𝐚𝐧 𝒙 ) 𝒂
𝒙
𝟏
𝟐𝛀(𝒂) = 𝐭𝐚𝐧−𝟏 𝒂 − 𝐭𝐚𝐧−𝟏 (𝒂). Hence,
𝟏 𝟏
𝛀(𝒂) = (𝐭𝐚𝐧−𝟏 𝒂 − 𝐭𝐚𝐧−𝟏 ( ))
𝟐 𝒂
Solution 2 by Rana Ranino-Setif-Algerie
𝟏 𝟏
𝒂 𝐬𝐢𝐧𝟐 𝒙 ⋅ 𝐭𝐚𝐧 ( ) 𝒙=
𝛀 ( 𝒂) = ∫ 𝒙𝟐 𝒅𝒙 =
𝒙
𝟏 𝟏 𝟏
𝒂 (𝟏 + 𝒙𝟐 )(𝐬𝐢𝐧𝟐 𝒙 𝐭𝐚𝐧 ( ) + 𝐬𝐢𝐧𝟐 ( 𝟐 ) 𝐭𝐚𝐧(𝒙𝟐 )
𝒙𝟐 𝒙
𝟏
𝒂 𝐬𝐢𝐧𝟐 ( 𝟐
𝟐 ) 𝐭𝐚𝐧 (𝒙 )
=∫ 𝒙 𝒅𝒙
𝟏
( 𝟏 + 𝒙 𝟐 ) (𝐬𝐢𝐧𝟐 𝒙 𝐭𝐚𝐧 ( 𝟏 ) + 𝐬𝐢𝐧𝟐 ( 𝟏 ) 𝐭𝐚𝐧(𝒙𝟐 ))
𝒂
𝒙𝟐 𝒙𝟐
𝒂
𝒅𝒙 −𝟏 −𝟏
𝟏
𝟐𝛀(𝒂) = ∫ = 𝐭𝐚𝐧 𝒂 − 𝐭𝐚𝐧 ( )
𝟏 𝟏 + 𝒙𝟐 𝒂
𝒂
𝝅
𝛀(𝒂) = 𝐭𝐚𝐧−𝟏 (𝒂) − 𝒔𝒈𝒏(𝒂)
𝟒
UP.397 Find:
𝒏
𝟐
𝛀 = 𝐥𝐢𝐦 ( √(𝟐𝒏 − 𝟏)‼) (𝒆𝑯𝟐𝒏+𝟐 −𝑯𝒏+𝟏 − 𝒆𝑯𝟐𝒏−𝑯𝒏 )
𝒏→∞

𝟏 𝟏 𝟏
Denote 𝒙𝒏 = 𝑯𝟐𝒏 − 𝑯𝒏 = 𝒏+𝟏 + 𝒏+𝟐 + ⋯ + 𝟐𝒏

www.ssmrmh.ro
𝟏
𝒙𝒏+𝟏 − 𝒙𝒏 = ; 𝒏 ∈ ℕ∗ , 𝐥𝐢𝐦 (𝒙𝒏+𝟏 − 𝒙𝒏 ) = 𝟎
(𝟐𝒏 + 𝟏)(𝟐𝒏 + 𝟐) 𝒏→∞
𝒏 𝟏
𝟏 𝟏 𝟏
𝐥𝐢𝐦 𝒙𝒏 = 𝐥𝐢𝐦 ∑ = ∫ 𝒅𝒙 = 𝐥𝐨𝐠 𝟐
𝒏→∞ 𝒏→∞ 𝒏 𝒌 𝟎 𝟏+𝒙
𝒌=𝟏 𝟏 +
𝒏
𝒆𝒙𝒏+𝟏 − 𝒆𝒙𝒏 𝒆𝒙𝒏 (𝒆𝒙𝒏+𝟏 −𝒙𝒏 − 𝟏) 𝒆𝒙𝒏+𝟏 −𝒙𝒏 − 𝟏
𝐥𝐢𝐦 = 𝐥𝐢𝐦 = 𝐥𝐢𝐦 ⋅ 𝐥𝐢𝐦 𝒆𝒙𝒏 = 𝟏 ⋅ 𝒆𝐥𝐨𝐠 𝟐 = 𝟐
𝒏→∞ 𝒙𝒏+𝟏 − 𝒙𝒏 𝒏→∞ 𝒙𝒏+𝟏 − 𝒙𝒏 𝒏→∞ 𝒙𝒏+𝟏 − 𝒙𝒏 𝒏→∞
Hence,
𝒏 𝟐
𝒏 √(𝟐𝒏 − 𝟏)‼
𝟐
𝛀 = 𝐥𝐢𝐦 (𝒆𝒙𝒏+𝟏 − 𝒆𝒙𝒏 ) ( √(𝟐𝒏 − 𝟏)‼) = 𝐥𝐢𝐦 ( ) ⋅ 𝒏𝟐 ⋅ (𝒆𝒙𝒏+𝟏 − 𝒆𝒙𝒏 ) =
𝒏→∞ 𝒏→∞ 𝒏
𝟒 𝒆𝒙𝒏+𝟏 − 𝒆𝒙𝒏 𝟐 𝟒 𝒏𝟐 𝟖 𝟏 𝟐
= 𝟐
⋅ 𝐥𝐢𝐦 ⋅ 𝒏 ( 𝒙 𝒏+𝟏 − 𝒙 𝒏 ) = 𝟐
⋅ 𝟐 ⋅ 𝐥𝐢𝐦 = 𝟐⋅ = 𝟐
𝒆 𝒏→∞ 𝒙𝒏+𝟏 − 𝒙𝒏 𝒆 𝒏→∞ (𝟐𝒏 + 𝟏)(𝟐𝒏 + 𝟐) 𝒆 𝟒 𝒆
Observation:
𝒏
√(𝟐𝒏 − 𝟏)‼ 𝒏 (𝟐𝒏 − 𝟏)‼
𝑪−𝑫 (𝟐𝒏 + !)‼ 𝒏𝒏
𝐥𝐢𝐦 = 𝐥𝐢𝐦 √ = 𝐥𝐢𝐦 ⋅ =
𝒏→∞ 𝒏 𝒏→∞ 𝒏𝒏 𝒏→∞ (𝒏 + !)𝒏+𝟏 (𝟐𝒏 − 𝟏)‼
(𝟐𝒏 − 𝟏)‼ (𝟐𝒏 + 𝟏) 𝟏 𝟏 𝟐

= 𝐥𝐢𝐦 ⋅ 𝒏 =𝟐⋅ =
𝒏→∞ (𝟐𝒏 − 𝟏)‼ (𝒏 + 𝟏) 𝟏 𝒆 𝒆
(𝟏 + 𝒏)
Solution 2 by Asmat Qatea-Afghanistan

(𝟐𝒏)! 𝟏 𝟏 𝒏 𝒏
(𝟐𝒏 − 𝟏)‼ = , 𝑯 ≅ 𝜸 + 𝐥𝐨𝐠 𝒏 + + 𝒐 ( ) ; 𝒏! ≅ √𝟐𝒏𝝅 ( )
𝟐𝒏 𝒏! 𝒏 𝟐𝒏 𝒏𝟐 𝒆
𝟐
(𝟐𝒏)!
𝒏 𝟏 𝟏 𝟏 𝟏
𝛀 = 𝟒 𝐥𝐢𝐦 ( √ ) (𝒆𝐥𝐨𝐠(𝟐𝒏+𝟐)+𝜸+𝟒𝒏+𝟒−(𝐥𝐨𝐠(𝒏+𝟏)+𝜸+𝟐𝒏+𝟐) − 𝒆𝐥𝐨𝐠(𝟐𝒏)+𝜸+𝟒𝒏−(𝐥𝐨𝐠 𝒏+𝜸+𝟐𝒏) )
𝒏→∞ 𝒏!
=
𝟐
𝟐𝒏
𝟐𝒏
𝒏 √𝟒𝝅𝒏 ( 𝒆 ) 𝐥𝐨𝐠 𝟐−
𝟏 𝟏
= 𝟒 𝐥𝐢𝐦 √ (𝒆 𝟒(𝒏+𝟏) − 𝒆𝐥𝐨𝐠 𝟐−𝟒𝒏 ) =
𝒏→∞ 𝒏 𝒏
√𝟐𝝅𝒏 ( 𝒆 )
( )
𝟏 𝟏 𝟏
−
= 𝐥𝐢𝐦 (
𝒏
√𝟐 ⋅ 𝒏𝟐)
(𝟐 ⋅ 𝒆 𝟒(𝒏+𝟏) − 𝟐 ⋅ 𝒆−𝟒𝒏 ) =
𝟐
𝒆 𝒏→∞

www.ssmrmh.ro
𝟏 𝟏
− +
𝟖 𝟏 𝒆 𝟒(𝒏+𝟏) 𝟒𝒏
−𝟏 𝟏 𝟏
= 𝟐 𝐥𝐢𝐦 (𝒏𝟐 ⋅ 𝒆−𝟒𝒏 ) ( ) ⋅ (− + )=
𝒆 𝒏→∞ 𝟏 𝟏 𝟒(𝒏 + 𝟏) 𝟒𝒏
− +
𝟒(𝒏 + 𝟏) 𝟒𝒏
𝟐 𝟐
𝟏 𝟐
= 𝐥𝐢𝐦 𝒏 ⋅ =
𝒆𝟐 𝒏→∞ 𝒏(𝒏 + 𝟏) 𝒆𝟐
Therefore,
𝒏
𝟐 𝟐
𝛀 = 𝐥𝐢𝐦 ( √(𝟐𝒏 − 𝟏)‼) (𝒆𝑯𝟐𝒏+𝟐 −𝑯𝒏+𝟏 − 𝒆𝑯𝟐𝒏 −𝑯𝒏 ) =
𝒏→∞ 𝒆𝟐
UP.398 If 𝟎 < 𝑎 ≤ 𝑏 then:
𝒂+𝒃
√𝒂𝒃 𝟐 𝒂+𝒃
𝟐 𝟐
∫ 𝒆−𝒙 𝒅𝒙 − ∫ 𝒆−𝒙 𝒅𝒙 ≥ 𝐭𝐚𝐧−𝟏(√𝒂𝒃) − 𝐭𝐚𝐧−𝟏 ( )
𝟎 𝟎 𝟐
𝒙 𝟐
Let be 𝒇: (𝟎, ∞) → ℝ, 𝒇(𝒙) = 𝐭𝐚𝐧−𝟏 𝒙 − ∫𝟎 𝒆−𝒕 𝒅𝒕 then,
𝟐
′(
𝟏 −𝒙𝟐
𝒆 𝒙 − ( 𝒙 𝟐 + 𝟏)
𝒇 𝒙) = − 𝒆 = ≥𝟎⇒
𝟏 + 𝒙𝟐 (𝟏 + 𝒙𝟐 )𝒆𝒙𝟐
𝟐
𝒇 −increasing. It is well-known that: 𝒆𝒙 ≥ 𝒙𝟐 + 𝟏, ∀𝒙 ≥ 𝟎
𝒂+𝒃 𝒂+𝒃
√𝒂𝒃 ≤ 𝟐
, 𝒇 −increasing, then 𝒇(√𝒂𝒃) ≤ 𝒇 ( 𝟐
)
𝒂+𝒃
√𝒂𝒃
−𝒕𝟐
𝒂+𝒃 𝟐 𝟐
𝐭𝐚𝐧−𝟏 (√𝒂𝒃) − ∫ 𝒆 𝒅𝒕 ≤ 𝐭𝐚𝐧−𝟏 ( )−∫ 𝒆−𝒕 𝒅𝒕
𝟎 𝟐 𝟎
Hence,
𝒂+𝒃
√𝒂𝒃
−𝒙𝟐
𝟐 𝟐 𝒂+𝒃
∫ 𝒆 𝒅𝒙 − ∫ 𝒆−𝒙 𝒅𝒙 ≥ 𝐭𝐚𝐧−𝟏 (√𝒂𝒃) − 𝐭𝐚𝐧−𝟏 ( )
𝟎 𝟎 𝟐
Equality holds for 𝒂 = 𝒃.
𝟐
Let 𝒈𝟏 (𝒙) = (𝟏 + 𝒙𝟐 )𝒆−𝒙 − 𝟏, 𝒙 ≥ 𝟎 then
𝟐 𝟐 𝟐
𝒈′𝟏 (𝒙) = 𝟐𝒙𝒆−𝒙 − 𝟐𝒙(𝟏 + 𝒙𝟐 )𝒆−𝒙 = −𝟐𝒙𝟑 𝒆−𝒙 < 0, ∀𝑥 > 0
𝟐 𝟏
⇒ 𝒈𝟏 (𝒙) < 𝒈𝟏 (𝟎) = 𝟏 ⇒ 𝒆−𝒙 < ; ∀𝒙 > 0
𝟏 + 𝒙𝟐

www.ssmrmh.ro
𝒙 𝒙
𝒅𝒕 𝟐
𝐋𝐞𝐭: 𝒈(𝒙) = ∫ 𝒆−𝒕 𝒅𝒕 − ∫ 𝟐
,𝒙 ≥ 𝟎
𝟎 𝟎 𝟏+𝒕
𝟐 𝟏
𝒈′ (𝒙) = 𝒆−𝒙 − 𝟏+𝒙𝟐 < 0; ∀𝑥 > 0 ⇒ 𝒈(𝒙) decreases on [𝟎, ∞)
𝒂+𝒃 𝒂+𝒃
≥ √𝒂𝒃, ∀ 𝒂, 𝒃 > 0 ⇒ 𝑔 ( ) ≤ 𝒈(√𝒂𝒃).
𝟐 𝟐
𝒂+𝒃 𝒂+𝒃
√𝒂𝒃 √𝒂𝒃
𝟐 𝟐 𝟐 𝒅𝒙 −𝒙𝟐
𝒅𝒙
⇒∫ 𝒆−𝒙 𝒅𝒙 − ∫ 𝟐
≤ ∫ 𝒆 𝒅𝒙 − ∫
𝟎 𝟎 𝟏+𝒙 𝟎 𝟎 𝟏 + 𝒙𝟐
𝒂+𝒃
√𝒂𝒃
𝟐 𝟐 𝟐 𝒂+𝒃
⇒∫ 𝒆−𝒙 𝒅𝒙 − ∫ 𝒆−𝒙 𝒅𝒙 ≤ 𝐭𝐚𝐧−𝟏 ( ) − 𝐭𝐚𝐧−𝟏 (√𝒂𝒃)
𝟎 𝟎 𝟐
𝒂+𝒃
√𝒂𝒃
𝟐 𝟐 𝟐 𝒂+𝒃
∫ 𝒆−𝒙 𝒅𝒙 − ∫ 𝒆−𝒙 𝒅𝒙 ≥ 𝐭𝐚𝐧−𝟏 (√𝒂𝒃) − 𝐭𝐚𝐧−𝟏 ( )
𝟎 𝟎 𝟐
𝒂+𝒃
𝟎 < 𝑎 ≤ 𝑏 ⇒ √𝒂𝒃 ≤
𝟐
𝒂+𝒃 𝒂+𝒃
𝟐 √𝒂𝒃 𝟐
−𝒙𝟐 −𝒙𝟐 𝟐
∫ 𝒆 𝒅𝒙 = ∫ 𝒆 𝒅𝒙 + ∫ 𝒆−𝒙 𝒅𝒙
𝟎 𝟎 √𝒂𝒃
𝒂+𝒃 𝒂+𝒃
√𝒂𝒃 𝟐 𝟐 √𝒂𝒃
𝟐 𝟐 𝟐 𝟐
∫ 𝒆−𝒙 𝒅𝒙 − ∫ 𝒆−𝒙 𝒅𝒙 = − ∫ 𝒆−𝒙 𝒅𝒙 = ∫ 𝒆−𝒙 𝒅𝒙
𝒂+𝒃
𝟎 𝟎 √𝒂𝒃 𝟐
√𝒂𝒃
−𝟏 (√𝒂𝒃) −𝟏
𝒂+𝒃 −𝟏 √𝒂𝒃 𝟏
𝐭𝐚𝐧 − 𝐭𝐚𝐧 ( ) = 𝐭𝐚𝐧 𝒙|𝒂+𝒃 = ∫ 𝒅𝒙
𝟐 𝒂+𝒃 𝟏 + 𝒙𝟐
𝟐
𝟐
So, we must prove that:

−𝒙𝟐
𝟏
∫ 𝒆 𝒅𝒙 ≥ ∫ 𝒅𝒙 ⇔
𝒂+𝒃 𝒂+𝒃 𝟏 + 𝒙𝟐
𝟐 𝟐
𝟐 𝟏 𝟐 𝟏 𝟏 𝟏 𝟐
−𝒆−𝒙 ≥ − 𝟐
⇔ 𝒆−𝒙 ≤ 𝟐 ⇔ 𝒙𝟐 ≤ 𝟐
⇔ 𝒆𝒙 ≥ 𝒙𝟐 + 𝟏 (𝐭𝐫𝐮𝐞)
𝟏+𝒙 𝒙 +𝟏 𝒆 𝟏+𝒙
𝒕=𝒙𝟐 𝟐 𝟏 𝟐
∵ 𝒆𝒕 ≥ 𝟏 + 𝒕; ∀𝒕 ≥ 𝟎 ⇒ 𝒆 𝒙 ≥ 𝟏 + 𝒙𝟐 ⇒ 𝟐
− 𝒆𝒙 ≥ 𝟎; ∀𝒙 ≥ 𝟎
𝟏+𝒙

www.ssmrmh.ro
𝒚
𝟏 𝟐 𝟏 𝟐
𝐋𝐞𝐭: 𝒇(𝒚) = ∫ ( 𝟐
− 𝒆−𝒙 ) 𝒅𝒙 𝐭𝐡𝐞𝐧 𝒇′ (𝒚) = 𝟐
− 𝒆−𝒚 ≥ 𝟎 ⇒ 𝒇 ↗
𝟎 𝟏 + 𝒙 𝟏 + 𝒚
𝒂+𝒃 𝒂+𝒃
𝐅𝐨𝐫 ≥ √𝒂𝒃 ⇒ 𝒇 ( ) ≥ 𝒇(√𝒂𝒃)
𝟐 𝟐
𝒂+𝒃
√𝒂𝒃
𝟐 𝟏 −𝒙𝟐
𝟏 𝟐
⇒∫ ( 𝟐
− 𝒆 ) 𝒅𝒙 ≥ ∫ ( 𝟐
− 𝒆−𝒙 ) 𝒅𝒙
𝟎 𝟏+𝒙 𝟎 𝟏+𝒙
𝒂+𝒃 𝒂+𝒃
−𝒙𝟐
𝟐
−𝒙𝟐
𝟏 𝟐 𝟏
⇒∫ 𝒆 𝒅𝒙 − ∫ 𝒆 𝒅𝒙 ≥ ∫ 𝒅𝒙 − ∫ 𝒅𝒙 ≥
𝟎 𝟎 𝟎 𝟏 + 𝒙𝟐 𝟎 𝟏 + 𝒙𝟐
𝒂+𝒃
≥ 𝐭𝐚𝐧−𝟏 (√𝒂𝒃) − 𝐭𝐚𝐧−𝟏 ( )
𝟐
UP.399 Find:
𝒕 𝒕
𝛀(𝒕) = 𝐥𝐢𝐦 ((𝚪(𝒙 + 𝟐))𝒙+𝟏 − (𝚪(𝒙 + 𝟏))𝒙 ) ⋅ 𝒙𝟏−𝒕 ; 𝒕 > 0
𝒙→∞

Solution by proposers
𝒕 𝟏
(𝚪(𝒙+𝟐))𝒙+𝟏 (𝚪(𝒙+𝟏))𝒙 𝟏
Denote: 𝒖(𝒙) = 𝒕 ; 𝐥𝐢𝐦 =
𝒙→∞ 𝒙 𝒆
(𝚪(𝒙+𝟏))𝒙
𝟏 𝒕 𝒕
(𝚪(𝒙 + 𝟐))𝒙+𝟏
𝒙 𝒙+𝟏 𝒕 𝟏 𝒕
𝐥𝐢𝐦 𝒖(𝒙) = 𝐥𝐢𝐦 ( ) ⋅( 𝟏
) ⋅ ( ) = 𝒕⋅𝒆 ⋅𝟏=𝟏
𝒙→∞ 𝒙→∞ 𝒙+𝟏 𝒙 𝒆
(𝚪(𝒙 + 𝟏))𝒙
𝒖(𝒙) − 𝟏
𝐥𝐢𝐦 = 𝟏;
𝒙→∞ 𝐥𝐨𝐠(𝒖(𝒙))
𝒕
𝚪(𝒙 + 𝟐) 𝟏 𝟏
𝐥𝐢𝐦 ( ) ⋅ 𝒕 = 𝐥𝐢𝐦 (𝒙 + 𝟏)𝒕 ⋅ 𝒕 =
𝒙→∞ 𝚪(𝒙 + 𝟏) 𝒙→∞
(𝚪(𝒙 + 𝟐))𝒙+𝟏 (𝚪(𝒙 + 𝟐))𝒙+𝟏
𝒕
𝒙+𝟏
= 𝐥𝐢𝐦 ( 𝟏
) = 𝒆𝒕
𝒙→∞
(𝚪(𝒙 + 𝟐))𝒙+𝟏
Hence,
𝒕
𝛀(𝒕) = 𝐥𝐢𝐦 ((𝚪(𝒙 + 𝟏))𝒙 ) (𝒖(𝒙) − 𝟏)𝒙𝟏−𝒕 =
𝒙→∞

www.ssmrmh.ro
𝟏 𝒕
(𝚪(𝒙 + 𝟏))𝒙 𝒖(𝒙) − 𝟏 𝒙 𝟏 𝒕
= 𝐥𝐢𝐦 ( ) ⋅ ⋅ 𝐥𝐨𝐠(𝒖(𝒙)) = 𝒕 ⋅ 𝟏 ⋅ 𝐥𝐨𝐠 𝒆𝒕 = 𝒕
𝐱→∞ 𝒙 𝐥𝐨𝐠 𝒖(𝒙) 𝒆 𝒆
𝝅
UP.400 If 𝟎 < 𝑎 ≤ 𝒃 < then:
𝟐
𝒃 (𝟑 − 𝐭𝐚𝐧 𝒙 )𝟐 𝒃 (𝟑 − 𝐜𝐨𝐭 𝒙)𝟐
∫ 𝒅𝒙 + ∫ 𝒅𝒙 ≥ 𝟏𝟖(𝒃 − 𝒂) + 𝟑 𝐥𝐨𝐠(𝐜𝐨𝐭 𝒂 ⋅ 𝐭𝐚𝐧 𝒃)
𝒂 𝐜𝐨𝐬 𝟐 (√𝐭𝐚𝐧 𝒙) 𝒂 𝐜𝐨𝐬 𝟐 (√𝐜𝐨𝐭 𝒙)

From MacLaurin series expansion for 𝒇(𝒙) = 𝐭𝐚𝐧 𝒙, we have that:

𝒙𝟑 𝟐𝒙𝟓 𝟏𝟕𝒙𝟕
𝐭𝐚𝐧 𝒙 ≥ 𝒙 + + +
𝟑 𝟏𝟓 𝟑𝟏𝟓
Hence,
𝒙𝟑 𝟐𝒙𝟓 𝟏𝟕𝒙𝟕
(𝟑 − 𝒙𝟐 ) 𝐭𝐚𝐧 𝒙 − 𝟑𝒙 ≥ (𝟑 − 𝒙𝟐 ) (𝒙 + + + ) − 𝟑𝒙 =
𝟑 𝟏𝟓 𝟑𝟏𝟓
𝟔𝒙𝟓 𝟓𝟏𝒙𝟕 𝒙𝟓 𝟐𝒙𝟕 𝟏𝟕𝒙𝟗
= (𝟑𝒙 + 𝒙𝟑 + + ) − (𝒙𝟑 + + + ) − 𝟑𝒙 =
𝟏𝟓 𝟑𝟏𝟓 𝟑 𝟏𝟓 𝟑𝟏𝟓
𝒙𝟓 (𝟐𝟏 + 𝟗𝒙𝟐 − 𝟏𝟕𝒙𝟒 )
= ≥𝟎
𝟑𝟏𝟓
Hence,
𝝅 𝟑𝒙 𝝅
(𝟑 − 𝒙𝟐 ) 𝐭𝐚𝐧 𝒙 ≥ 𝟑𝒙, ∀𝒙 ∈ (𝟎, ) ⟺ 𝐭𝐚𝐧 𝒙 ≥ , ∀𝒙 ∈ (𝟎, )
𝟐 𝟑 − 𝒙𝟐 𝟐
Integrating (𝟏) we have:
𝒙 𝒙
𝟑𝒕 𝟑
∫ 𝐭𝐚𝐧 𝒕 𝒅𝒕 ≥ ∫ 𝒅𝒕 ⇒ − 𝐥𝐨𝐠 ( 𝐜𝐨𝐬 𝒙 ) ≥ − (𝐥𝐨𝐠(𝟑 − 𝒙𝟐 ) − 𝐥𝐨𝐠 𝟑)
𝟎 𝟎 𝟑 − 𝒕𝟐 𝟐
𝟑
𝟐
𝟑 − 𝒙𝟐 𝝅 (𝟑 − 𝒙𝟐 )𝟐 𝝅
𝐜𝐨𝐬 𝒙 ≤ ( ) , ∀𝒙 ∈ (𝟎, ) ⟺ 𝟐
≥ 𝟗 + 𝟑𝒙𝟐 , ∀𝒙 ∈ (𝟎, ) ; (𝟐)
𝟑 𝟐 𝐜𝐨𝐬 𝒙 𝟐
Now, in (2) putting 𝒙 → √𝐭𝐚𝐧 𝒙 and 𝒙 → √𝐜𝐨𝐭 𝒙, we get:

www.ssmrmh.ro
(𝟑 − 𝐭𝐚𝐧 𝒙)𝟐
≥ 𝟗 + 𝟑 𝐭𝐚𝐧 𝒙
𝐜𝐨𝐬 𝟐 (√𝐭𝐚𝐧 𝒙)
⇒
(𝟑 − 𝐜𝐨𝐭 𝒙)𝟐
≥ 𝟗 + 𝟑 𝐜𝐨𝐭 𝒙
{ 𝐜𝐨𝐬 𝟐 (√𝐜𝐨𝐭 𝒙)
(𝟑 − 𝐭𝐚𝐧 𝒙)𝟐 (𝟑 − 𝐜𝐨𝐭 𝒙)𝟐
+ ≥ 𝟏𝟖 + 𝟑(𝐭𝐚𝐧 𝒙 + 𝐜𝐨𝐭 𝒙)
𝐜𝐨𝐬 𝟐 (√𝐭𝐚𝐧 𝒙) 𝐜𝐨𝐬 𝟐 (√𝐜𝐨𝐭 𝒙)
Hence,
𝒃 (𝟑 − 𝐭𝐚𝐧 𝒙)𝟐 𝒃 (𝟑 − 𝐜𝐨𝐭 𝒙)𝟐
∫ 𝒅𝒙 + ∫ 𝒅𝒙 ≥ 𝟏𝟖(𝒃 − 𝒂) + 𝟑 𝐥𝐨𝐠(𝐜𝐨𝐭 𝒂 ⋅ 𝐭𝐚𝐧 𝒃)
𝒂 𝐜𝐨𝐬 𝟐 (√𝐭𝐚𝐧 𝒙) 𝒂 𝐜𝐨𝐬 𝟐 (√𝐜𝐨𝐭 𝒙)
𝝅
UP.401 If 𝟎 < 𝑎 < 𝑏 < then:
𝟐
𝒃
𝐜𝐨𝐬 𝒙 𝐬𝐢𝐧 𝒙 𝐬𝐢𝐧 𝒙 ⋅ 𝐭𝐚𝐧 𝒙 𝐜𝐨𝐬 𝒙 ⋅ 𝐜𝐨𝐭 𝒙 𝟓
∫ ( + + + ) 𝒅𝒙 ≤ 𝐥𝐨𝐠(𝐜𝐨𝐭 𝒂 ⋅ 𝐭𝐚𝐧 𝒃)
𝒂 𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝐭𝐚𝐧(𝐜𝐨𝐬 𝒙) 𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝐭𝐚𝐧(𝐜𝐨𝐬 𝒙) 𝟑
𝐜𝐨𝐬 𝒙 𝐬𝐢𝐧 𝒙 𝐬𝐢𝐧 𝒙 ⋅ 𝐭𝐚𝐧 𝒙 𝐜𝐨𝐬 𝒙 ⋅ 𝐜𝐨𝐭 𝒙
+ + + =
𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝐭𝐚𝐧(𝐜𝐨𝐬 𝒙) 𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝐭𝐚𝐧(𝐜𝐨𝐬 𝒙)
𝐜𝐨𝐬 𝒙 𝐬𝐢𝐧 𝒙 𝐬𝐢𝐧𝟐 𝒙 𝐜𝐨𝐬 𝟐 𝒙
= + + + =
𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝐭𝐚𝐧(𝐜𝐨𝐬 𝒙) 𝐜𝐨𝐬 𝒙 ⋅ 𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝐬𝐢𝐧 𝒙 ⋅ 𝐭𝐚𝐧(𝐜𝐨𝐬 𝒙)
𝐬𝐢𝐧 𝒙 𝐜𝐨𝐬 𝒙
=( + ) (𝐭𝐚𝐧 𝒙 + 𝐜𝐨𝐭 𝒙); (𝟏)
𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝐭𝐚𝐧(𝐜𝐨𝐬 𝒙)
Now, from Maclaurin series expansion for 𝒇(𝒙) = 𝐭𝐚𝐧 𝒙, we have that:
𝒙𝟑 𝟐𝒙𝟓 𝟏𝟕𝒙𝟕
𝐭𝐚𝐧 𝒙 ≥ 𝒙 + + +
𝟑 𝟏𝟓 𝟑𝟏𝟓
Hence,
𝒙𝟑 𝟐𝒙𝟓 𝟏𝟕𝒙𝟕
(𝟑 − 𝒙𝟐 ) 𝐭𝐚𝐧 𝒙 − 𝟑𝒙 ≥ (𝟑 − 𝒙𝟐 ) (𝒙 + + + ) − 𝟑𝒙 =
𝟑 𝟏𝟓 𝟑𝟏𝟓
𝟔𝒙𝟓 𝟓𝟏𝒙𝟕
𝟑 𝟑
𝒙𝟓 𝟐𝒙𝟕 𝟏𝟕𝒙𝟗
= (𝟑𝒙 + 𝒙 + + ) − (𝒙 + + + ) − 𝟑𝒙 =
𝟏𝟓 𝟑𝟏𝟓 𝟑 𝟏𝟓 𝟑𝟏𝟓
𝒙𝟓 (𝟐𝟏 + 𝟗𝒙𝟐 − 𝟏𝟕𝒙𝟒 )
= ≥𝟎
𝟑𝟏𝟓
www.ssmrmh.ro
Hence,
𝝅
(𝟑 − 𝒙𝟐 ) 𝐭𝐚𝐧 𝒙 ≥ 𝟑𝒙, ∀𝒙 ∈ (𝟎, )
𝟐
For 𝒙 → 𝐬𝐢𝐧 𝒙 and 𝒙 → 𝐜𝐨𝐬 𝒙 it follows that:
𝐬𝐢𝐧 𝒙 𝐜𝐨𝐬 𝒙 𝟑 − 𝐬𝐢𝐧𝟐 𝒙 𝟑 − 𝐜𝐨𝐬 𝟐 𝒙 𝟓 𝝅
+ ≤ + = , ∀𝒙 ∈ (𝟎, ) ; (𝟐)
𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝐭𝐚𝐧(𝐜𝐨𝐬 𝒙) 𝟑 𝟑 𝟑 𝟐
𝒃 𝒃 𝒃 𝒃
𝐜𝐨𝐬 𝒙 𝐬𝐢𝐧 𝒙 𝐬𝐢𝐧 𝒙 ⋅ 𝐭𝐚𝐧 𝒙 𝐜𝐨𝐬 𝒙 ⋅ 𝐜𝐨𝐭 𝒙
∫ 𝒅𝒙 + ∫ 𝒅𝒙 + ∫ 𝒅𝒙 + ∫ 𝒅𝒙 =
𝒂 𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝒂 𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝒂 𝐭𝐚𝐧 (𝐬𝐢𝐧 𝒙) 𝒂 𝐭𝐚𝐧(𝐜𝐨𝐬 𝒙)
𝒃
𝐬𝐢𝐧 𝒙 𝐜𝐨𝐬 𝒙
=∫ ( + ) (𝐭𝐚𝐧 𝒙 + 𝐜𝐨𝐭 𝒙)𝒅𝒙 ≤
𝒂 𝐭𝐚𝐧(𝐬𝐢𝐧 𝒙) 𝐭𝐚𝐧(𝐜𝐨𝐬 𝒙)
𝟓 𝒃 𝟓
≤ ∫ (𝐭𝐚𝐧 𝒙 + 𝐜𝐨𝐭 𝒙) 𝒅𝒙 = 𝐥𝐨𝐠(𝐜𝐨𝐭 𝒂 ⋅ 𝐭𝐚𝐧 𝒃)
𝟑 𝒂 𝟑
𝟏
UP.402 If < 𝑎 ≤ 𝑏 < 𝑒 then:
𝒆
𝒂+𝒃 𝒂+𝒃
𝐬𝐢𝐧−𝟏 (𝐥𝐨𝐠(√𝒂𝒃)) − 𝐬𝐢𝐧−𝟏 (𝐥𝐨𝐠 ( )) ≤ √𝟏 − 𝐥𝐨𝐠 𝟐 ( ) − √𝟏 − 𝐥𝐨𝐠 𝟐 (√𝒂𝒃)
𝟐 𝟐

𝟏
Let be 𝒇: [ , 𝒆] → ℝ, 𝒇(𝒙) = 𝐬𝐢𝐧−𝟏 (𝐥𝐨𝐠 𝒙) − √𝟏 − 𝐥𝐨𝐠 𝟐 𝒙 then,
𝒆
𝟐
𝟏 − 𝒙 𝐥𝐨𝐠 𝒙 𝟏 − 𝐥𝐨𝐠 𝒙 𝟏
𝒇′ (𝒙) = + = ≥ 𝟎, ∀𝒙 ∈ [ , 𝒆]
𝒙√𝟏 − 𝐥𝐨𝐠 𝟐 𝒙 𝟐√𝟏 − 𝐥𝐨𝐠 𝟐 𝒙 𝒙√𝟏 − 𝐥𝐨𝐠 𝟐 𝒙 𝒆
𝒂+𝒃 𝒂+𝒃
𝒇 −inreasing and from √𝒂𝒃 ≤ ⇒ 𝒇(√𝒂𝒃) ≤ 𝒇 ( ).
𝟐 𝟐
Hence,
𝒂+𝒃 𝒂+𝒃
𝐬𝐢𝐧−𝟏 (𝐥𝐨𝐠(√𝒂𝒃)) + √𝟏 − 𝐥𝐨𝐠 𝟐 (√𝒂𝒃) ≤ 𝐬𝐢𝐧−𝟏 (𝐥𝐨𝐠 ( )) + √𝟏 − 𝐥𝐨𝐠 𝟐 ( )
𝟐 𝟐
And thus,

www.ssmrmh.ro
𝒂+𝒃 𝒂+𝒃
𝟐 𝟐

Consider 𝒇: (−𝟏, 𝟏) → ℝ, 𝒇(𝒙) = 𝐬𝐢𝐧−𝟏 𝒙 + √𝟏 − 𝒙𝟐, with derivatives
𝟏−𝒙
𝒇′ (𝒙) = > 0; ∀𝑥 ∈ (−𝟏, 𝟏)
√𝟏 − 𝒙𝟐
𝒂+𝒃
We have 𝐥𝐨𝐠 √𝒂𝒃 < 𝐥𝐨𝐠 . Using monotony for 𝒇, we obtain
𝒂
𝒂+𝒃 𝒂+𝒃
𝐬𝐢𝐧−𝟏 (𝐥𝐨𝐠(√𝒂𝒃)) + √𝟏 − 𝐥𝐨𝐠 𝟐 (√𝒂𝒃) ≤ 𝐬𝐢𝐧−𝟏 (𝐥𝐨𝐠 ( )) + √𝟏 − 𝐥𝐨𝐠 𝟐 ( )
𝟐 𝟐
That is another form for
𝒂+𝒃 𝒂+𝒃
𝟐 𝟐

UP.403 If 𝟎 < 𝑎 ≤ 𝑏 then:
𝒃 𝒃
√𝟏𝟒
∫ ∫ √𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 𝒅𝒙𝒅𝒚 ≥ (𝒃 + 𝒂)(𝒃 − 𝒂)𝟐
𝒂 𝒂 𝟐
Solution 1by proposer
𝟕(𝒙 − 𝒚)𝟐 + (𝟕𝒙 + 𝟐𝟏𝒚)𝟐 = 𝟕(𝒙𝟐 − 𝟐𝒙𝒚 + 𝒚𝟐 ) + 𝟒𝟗𝒙𝟐 + 𝟒𝟒𝟏𝒚𝟐 + 𝟐𝟗𝟒𝒙𝒚 =
= 𝟓𝟔𝒙𝟐 − 𝟏𝟒𝒙𝒚 + 𝟕𝒚𝟐 + 𝟒𝟒𝟏𝒚𝟐 + 𝟐𝟗𝟒𝒙𝒚 =
= 𝟓𝟔𝒙𝟐 + 𝟐𝟖𝟎𝒙𝒚 + 𝟒𝟒𝟖𝒚𝟐 = 𝟓𝟔(𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 )
Hence,
𝟏
𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 = (𝟕(𝒙 − 𝒚)𝟐 + (𝟕𝒙 + 𝟐𝟏𝒚)𝟐 ) ≥
𝟓𝟔
𝟏 𝟒𝟗 𝟕
≥ (𝟕𝒙 + 𝟐𝟏𝒚)𝟐 = (𝒙 + 𝟑𝒚)𝟐 = (𝒙 + 𝟑𝒚)𝟐
𝟓𝟔 𝟓𝟔 𝟖

www.ssmrmh.ro
So, we get:
√𝟕
√𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 ≥ (𝒙 + 𝟑𝒚)
𝟐√𝟐
Therefore,
𝒃 𝒃
√𝟕 ⋅ √𝟐 𝒃 𝒃
∫ ∫ √𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 𝒅𝒙𝒅𝒚 ≥ ∫ ∫ (𝒙 + 𝟑𝒚) 𝒅𝒙𝒅𝒚 =
𝒂 𝒂 𝟒 𝒂 𝒂
𝒃 𝒃 𝒃
√𝟏𝟒 √𝟏𝟒 𝒙𝟐
= (𝒃 − 𝒂) (∫ 𝒙 𝒅𝒙 + ∫ 𝟑𝒚 𝒅𝒚) = (𝒃 − 𝒂) ⋅ 𝟒 ⋅ | =
𝟒 𝒂 𝒂 𝟒 𝟐 𝒂
√𝟏𝟒 √𝟏𝟒
= (𝒃 − 𝒂)(𝒃𝟐 − 𝒂𝟐 ) = (𝒃 + 𝒂)(𝒃 − 𝒂)𝟐
𝟐 𝟐
Solution 2 by George Florin Șerban-Romania
𝑨𝑮𝑴 𝒚≥𝒙
𝟏𝟒
If 𝒙 ≥ 𝒚, 𝒙𝟓 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 ≥ 𝟏𝟒 √𝒙𝟐 ⋅ 𝒙𝟓 𝒚𝟓 ⋅ 𝒚𝟏𝟔 = 𝟏𝟒𝒚√𝒙𝒚 ≥ 𝟏𝟒𝒚𝟐
Equality holds for 𝒙 = 𝒚.
⇒ √𝒙𝟓 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 ≥ √𝟏𝟒𝒚𝟐 = 𝒚√𝟏𝟒

𝒃 𝒃 𝒃 𝒃 𝒃
𝒚𝟐
∫ ∫ √𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 𝒅𝒙𝒅𝒚 ≥ ∫ ∫ √𝟏𝟒𝒚 𝒅𝒙𝒅𝒚 = (𝒃 − 𝒂)√𝟏𝟒 ⋅ | =
𝒂 𝒂 𝒂 𝒂 𝟐 𝒂
(𝒃 − 𝒂)𝟐 (𝒃 + 𝒂)√𝟏𝟒
=
𝟐
If 𝒙 ≤ 𝒚, 𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 ≥ 𝒙√𝟏𝟒 ⇒
𝒃 𝒃 𝒃 𝒃 𝒃
𝒙𝟐
∫ ∫ √𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 𝒅𝒙𝒅𝒚 ≥ ∫ ∫ √𝟏𝟒𝒙 𝒅𝒙𝒅𝒚 = (𝒃 − 𝒂)√𝟏𝟒 ⋅ | =
𝒂 𝒂 𝒂 𝒂 𝟐 𝒂
(𝒃 − 𝒂)𝟐 (𝒃 + 𝒂)√𝟏𝟒
=
𝟐
Therefore,
𝒃 𝒃 (𝒃 − 𝒂)𝟐 (𝒃 + 𝒂)√𝟏𝟒
∫ ∫ √𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 𝒅𝒙𝒅𝒚 ≥
𝒂 𝒂 𝟐

www.ssmrmh.ro
We have:
𝟐
𝟐𝒙 + 𝟓𝒚
𝟐√𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 = √(𝟐𝒙 + 𝟓𝒚)𝟐 + 𝟕𝒚𝟐 = √𝟕 ⋅ √𝟕 ( ) + 𝒚𝟐 =
𝟕
𝟕 𝟐𝒙 + 𝟓𝒚 𝟐 𝟐𝒙 + 𝟓𝒚 𝟐 𝑪𝑩𝑺
= √ √(𝟏
⏟+ ⋯ + 𝟏) (( ) +⋯+ ( ) ) + 𝒚𝟐 ≥
𝟖 𝟖−𝒕𝒊𝒎𝒆𝒔
⏟ 𝟕 𝟕
𝟕−𝒕𝒊𝒎𝒆𝒔
√𝟏𝟒 𝟐𝒙 + 𝟓𝒚 𝟐𝒙 + 𝟓𝒚
≥ (( )+ ⋯+ ( ) + 𝒚)
𝟒 ⏟ 𝟕 𝟕
𝟕−𝒕𝒊𝒎𝒆𝒔
√𝟏𝟒
⇒ √𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 ≥ (𝒙 + 𝟑𝒚); ∀𝒙, 𝒚 ≥ 𝟎
𝟒
𝒃 𝒃
√𝟏𝟒 𝒃 𝒃
∫ ∫ √𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 𝒅𝒙𝒅𝒚 ≥ ∫ ∫ (𝒙 + 𝟑𝒚) 𝒅𝒙𝒅𝒚 =
𝒂 𝒂 𝟒 𝒂 𝒂
√𝟏𝟒 𝒃𝟐 𝒂𝟐 (𝒃 − 𝒂)𝟐 (𝒃 + 𝒂)√𝟏𝟒

= (𝟏 + 𝟑) ( − ) (𝒃 − 𝒂) =
𝟒 𝟐 𝟐 𝟐
Therefore,
𝒃 𝒃 (𝒃 − 𝒂)𝟐 (𝒃 + 𝒂)√𝟏𝟒
∫ ∫ √𝒙𝟐 + 𝟓𝒙𝒚 + 𝟖𝒚𝟐 𝒅𝒙𝒅𝒚 ≥
𝒂 𝒂 𝟐
UP.404 If 𝟎 < 𝑎 ≤ 𝑏 < 1 then:
𝟏 − 𝒂 + 𝒃 − 𝒂𝒃 𝟑√𝟑 𝟐
𝐥𝐨𝐠 ( )≥ (𝒃 − 𝒂𝟐 )
𝟏 + 𝒂 − 𝒃 − 𝒂𝒃 𝟐
Inequality can be written:
(𝟏 + 𝒃)(𝟏 − 𝒂) 𝟑√𝟑 𝟐
𝐥𝐨𝐠 ( )≥ ( 𝒃 − 𝒂𝟐 )
(𝟏 + 𝒂)(𝟏 − 𝒃) 𝟐
𝟏 𝟏−𝒃 𝟏+𝒂 𝟑√𝟑 𝟐

− 𝐥𝐨𝐠 ( ⋅ )≥ (𝒃 − 𝒂𝟐 )
𝟐 𝟏+𝒃 𝟏−𝒂 𝟒

www.ssmrmh.ro
𝟏 𝟏−𝒃 𝟏+𝒂 𝟑√𝟑 𝟐
− (𝐥𝐨𝐠 ( ) − 𝐥𝐨𝐠 ( )) ≥ (𝒃 − 𝒂𝟐 )
𝟐 𝟏+𝒃 𝟏−𝒂 𝟒
𝒃
𝟏 𝟏 − 𝒙 𝒃 𝟑√𝟑 𝒙𝟐
− ⋅ 𝐥𝐨𝐠 ( )| ≥ ⋅ |
𝟐 𝟏+𝒙 𝒂 𝟑 𝟐 𝒂
𝒃
𝟏 𝟑√𝟑 𝒃
∫ 𝒅𝒙 ≥ ∫ 𝒙 𝒅𝒙
𝒂 𝟏 − 𝒙𝟐 𝟐 𝒂
𝟏 𝟑√𝟑
It is enough to prove that ∀𝒙 ∈ (𝟎, 𝟏): ≥ 𝒙⇔
𝟏−𝒙𝟐 𝟐
𝟐
𝟏 − 𝒙𝟐 ≤ ⇔ 𝟑√𝟑𝒙(𝟏 − 𝒙𝟐 ) ≤ 𝟐 ⇔ 𝟑√𝟑𝒙(𝒙𝟐 − 𝟏) + 𝟐 ≥ 𝟎
𝟑√𝟑𝒙
𝟑√𝟑𝒙𝟑 − 𝟑√𝟑𝒙 + 𝟐 ≥ 𝟎
𝟑√𝟑𝒙𝟑 + 𝟔𝒙𝟐 − 𝟔𝒙𝟐 − 𝟒𝒙√𝟑 + 𝒙√𝟑 + 𝟐 ≥ 𝟎
(𝟑𝒙𝟐 − 𝟐𝒙√𝟑 + 𝟏)(𝒙√𝟑 + 𝟐) ≥ 𝟎
𝟐
(𝒙√𝟑 − 𝟏) (𝒙√𝟑 + 𝟐) ≥ 𝟎 true.
(𝟏 + 𝒃)(𝟏 − 𝒂) 𝟑√𝟑 𝟐
(∗) ↔ 𝐥𝐨𝐠 ( )≥ (𝒃 − 𝒂𝟐 ) ↔
(𝟏 + 𝒂)(𝟏 − 𝒃) 𝟐
𝟏+𝒃 𝟏−𝒃 𝒃𝟐 𝒂𝟐
𝐥𝐨𝐠 ( ) − 𝐥𝐨𝐠 ( ) ≥ 𝟑√𝟑 ( − )
𝟏+𝒂 𝟏−𝒂 𝟐 𝟐
𝒃 𝒃
𝟏 𝟏
↔ ∫( + ) 𝒅𝒙 ≥ ∫ 𝟑√𝟑𝒙𝒅𝒙
𝟏+𝒙 𝟏−𝒙
𝒂 𝒂
𝟏 𝟏
→ 𝑰𝒕′ 𝒔 𝒔𝒖𝒇𝒇𝒊𝒄𝒆𝒔 𝒕𝒐 𝒑𝒓𝒐𝒗𝒆 ∶
+ ≥ 𝟑√𝟑𝒙, ∀𝒙 ∈ (𝟎, 𝟏)
𝟏+𝒙 𝟏−𝒙
𝟏 𝟏 𝟐 𝟖
+ ≥ 𝟑√𝟑𝒙 ↔ ≥ 𝒙( 𝟏 − 𝒙𝟐 ) ↔ ≥ 𝟐𝒙𝟐 (𝟏 − 𝒙𝟐 )𝟐
𝟏+𝒙 𝟏−𝒙 𝟑√𝟑 𝟐𝟕
𝟑
𝟐( 𝟐 )𝟐
𝟐𝒙𝟐 + (𝟏 − 𝒙𝟐 ) + (𝟏 − 𝒙𝟐 ) 𝟖
𝑩𝒚 𝑨𝑴 − 𝑮𝑴, 𝒘𝒆 𝒉𝒂𝒗𝒆 ∶ 𝟐𝒙 𝟏 − 𝒙 ≤( ) =
𝟑 𝟐𝟕
𝟏 − 𝒂 + 𝒃 − 𝒂𝒃 𝟑√𝟑 𝟐
𝐓𝐡𝐞𝐫𝐞𝐟𝐨𝐫𝐞, 𝐥𝐨𝐠 ( )≥ (𝒃 − 𝒂𝟐 )
𝟏 + 𝒂 − 𝒃 − 𝒂𝒃 𝟐

www.ssmrmh.ro
𝟏 − 𝒂 + 𝒃 − 𝒂𝒃 𝟑√𝟑 𝟐
𝐅𝐨𝐫 𝟎 < 𝒂 ≤ 𝒃 < 𝟏, 𝐥𝐨𝐠 ( )≥ (𝒃 − 𝒂𝟐 ); (𝟏)
𝟏 + 𝒂 − 𝒃 − 𝒂𝒃 𝟐
Is clearly holds for 𝒂 = 𝒃.
Let us assume 𝟎 < 𝒂 < 𝒃 < 𝟏, then we can rewrite (1) as
𝟏+𝒃 𝟏+𝒂
𝐥𝐨𝐠 (𝟏 − 𝒃) − 𝐥𝐨𝐠 (𝟏 − 𝒂) 𝟑√𝟑
≥ ; ( 𝟐)
𝒃𝟐 − 𝒂𝟐 𝟐
𝟏+𝒙
𝐋𝐞𝐭 𝒇(𝒙) = 𝐥𝐨𝐠 ( ) , 𝟎 < 𝒙 < 𝟏 𝐚𝐧𝐝 𝒈(𝒙) = 𝒙𝟐 , 𝟎 < 𝒙 < 𝟏.
𝟏−𝒙
For 𝟎 < 𝒂 < 𝒃 < 𝟏, by the Cauchy’s mean value theorem,
𝒇(𝒃) − 𝒇(𝒂) 𝒇′ (𝒄) 𝟏 𝟏 𝟏 𝟏
𝐟𝐨𝐫 𝐬𝐨𝐦𝐞 𝒄 ∈ (𝒂, 𝒃) : = ′ = ( + )= ; (𝟑)
𝒈(𝒃) − 𝒈(𝒂) 𝒈 (𝒄) 𝟐𝒄 𝟏 + 𝒄 𝟏 − 𝒄 𝒄(𝟏 − 𝒄𝟐 )
Let 𝒉(𝒙) = 𝒙(𝟏 − 𝒙𝟐 ), 𝟎 < 𝒙 < 𝟏, then 𝒉′ (𝒙) = 𝟏 − 𝟑𝒙𝟐 = (𝟏 − √𝟑𝒙)(𝟏 + √𝟑𝒙)
𝟏 𝟏 𝟐
𝒉′ (𝒙) > 𝟎 if 𝟎 < 𝒙 < ⇒ 𝒉 ( 𝒙) ≤ 𝟏 = 𝟑√𝟑 , ∀𝒙 ∈ (𝟎, 𝟏)
√𝟑 √𝟑(𝟏−𝟑)
𝟐 𝟏 𝟐
⇒ 𝒉(𝒄) ≤ ⇒ ≥ , ∀𝒄 ∈ (𝟎, 𝟏); (𝟒)
𝟑√𝟑 𝒄(𝟏 − 𝒄𝟐 ) 𝟑√𝟑
From (3),(4) we get (2).
Solution 4 by Hikmat Mammadov-Azerbaijan
𝒂 ≤ 𝒃 ⇒ −𝒂 ≥ −𝒃 − 𝒂 ≥ 𝟏 − 𝒃, 𝒃 ≥ 𝒂 ⇒ 𝟏 + 𝒃 ≥ 𝟏 + 𝒂
⇒ (𝟏 − 𝒂)(𝟏 + 𝒃) ≥ (𝟏 − 𝒃)(𝟏 + 𝒂) > 𝟎
(𝟏 − 𝒂)(𝟏 + 𝒃) 𝟏 − 𝒂 + 𝒃 − 𝒂𝒃 𝟏 − 𝒂 + 𝒃 − 𝒂𝒃
⇒ ≥𝟏⇒ ≥ 𝟏 ⇒ 𝐥𝐨𝐠 ( )≥𝟎
(𝟏 + 𝒂)(𝟏 − 𝒃) 𝟏 + 𝒂 − 𝒃 − 𝒂𝒃 𝟏 + 𝒂 − 𝒃 − 𝒂𝒃
𝟏 ≥ 𝒃𝟐 − 𝒂𝟐 ≥ 𝟎
𝟑√𝟑 𝟑√𝟑 𝟐
> (𝒃 − 𝒂𝟐 ) ≥ 𝟎
𝟐 𝟐
𝟑√𝟑 𝟐 𝟑√𝟑 𝟏 − 𝒂 + 𝒃 − 𝒂𝒃
𝐖𝐡𝐞𝐧: (𝒃 − 𝒂𝟐 ) ≅ 𝐥𝐨𝐠 ( )≅∞
𝟐 𝟐 𝟏 + 𝒂 − 𝒃 − 𝒂𝒃
𝟏 − 𝒂 + 𝒃 − 𝒂𝒃 𝟑√𝟑 𝟐
𝐓𝐡𝐞𝐫𝐞𝐟𝐨𝐫𝐞, 𝐥𝐨𝐠 ( )≥ (𝒃 − 𝒂𝟐 )
𝟏 + 𝒂 − 𝒃 − 𝒂𝒃 𝟐

www.ssmrmh.ro
UP.405 Find:
𝝅
𝒎 𝝅
𝛀(𝒎) = ∫ 𝐬𝐢𝐧(𝒎𝒙) ⋅ 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( ) 𝐭𝐚𝐧 𝒙) 𝒅𝒙 ; 𝒎 > 2
𝟎 𝒎
Solution by proposers
𝝅
Let 𝒙 = 𝒎 − 𝒚, 𝒅𝒙 = −𝒅𝒚 then,
𝟎
𝝅 𝝅 𝝅
𝛀(𝒎) = ∫ 𝐬𝐢𝐧 (𝒎 ( − 𝒚)) 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( ) 𝒕𝒂𝒏 ( − 𝒚)) (−𝒅𝒚) =
𝝅 𝒎 𝒎 𝒎
𝒎
𝝅 𝝅
𝒎 𝝅 𝐭𝐚𝐧 (𝒎) − 𝐭𝐚𝐧 𝒚
= ∫ 𝐬𝐢𝐧(𝝅 − 𝒎𝒚) 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( ) ⋅ ) 𝒅𝒚 =
𝟎 𝒎 𝟏 + 𝐭𝐚𝐧 ( 𝝅 ) 𝐭𝐚𝐧 𝒚
𝒎
𝝅 𝝅 𝝅 𝝅
𝒎 𝟏 + 𝐭𝐚𝐧 (𝒎) 𝐭𝐚𝐧 𝒚 + 𝐭𝐚𝐧𝟐 (𝒎) − 𝐭𝐚𝐧 (𝒎) 𝐭𝐚𝐧 𝒚
= ∫ 𝐬𝐢𝐧(𝒎𝒚) 𝐥𝐨𝐠 ( 𝝅 ) 𝒅𝒚 =
𝟎 𝟏 + 𝐭𝐚𝐧 ( ) 𝐭𝐚𝐧 𝒚
𝒎
𝝅 𝝅
𝒎 𝟏 + 𝐭𝐚𝐧𝟐 (𝒎)
= ∫ 𝐬𝐢𝐧(𝒎𝒙) 𝐥𝐨𝐠 ( 𝝅 ) 𝒅𝒙 =
𝟎 𝟏 + 𝐭𝐚𝐧 (𝒎) 𝐭𝐚𝐧 𝒙
𝝅 𝝅
𝒎 𝝅 𝟐
𝒎 𝝅
= ∫ 𝐬𝐢𝐧(𝒎𝒙) 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( )) 𝒅𝒙 − ∫ 𝐬𝐢𝐧(𝒎𝒙) 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( ) 𝐭𝐚𝐧 𝒙) 𝒅𝒙 =
𝟎 𝒎 𝟎 𝒎
𝝅
𝟐
𝝅 𝒎
= 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( )) ∫ 𝐬𝐢𝐧(𝒎𝒙) 𝒅𝒙 − 𝛀(𝒎)
𝒎 𝟎
Hence,
𝝅
𝝅 𝟏 𝒎
𝟐𝛀(𝒎) = 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧𝟐 ( )) ⋅ (− ) 𝐜𝐨𝐬(𝒎𝒙)| =
𝒎 𝒎 𝟎
𝟏 𝟏 𝝅
= 𝐥𝐨𝐠 ( 𝝅 ) (− ) (𝐜𝐨𝐬 (𝒎 ⋅ ) − 𝐜𝐨𝐬 𝟎) =
𝐜𝐨𝐬 𝟐 (𝒎) 𝒎 𝒎
𝟏 𝝅 𝟏 𝟏 𝝅
− 𝐥𝐨𝐠 (𝐜𝐨𝐬 ( )) ⋅ (− ) (𝐜𝐨𝐬 𝝅 − 𝐜𝐨𝐬 𝟎) = 𝐥𝐨𝐠 (𝐜𝐨𝐬 ( )) ⋅ (−𝟐)
𝟐 𝒎 𝒎 𝟐𝒎 𝒎

www.ssmrmh.ro
Therefore,
𝟏 𝝅
𝛀(𝒎) = − 𝐥𝐨𝐠 (𝐜𝐨𝐬 ( ))
𝟐𝒎 𝒎
𝒙=𝟎 𝝅
𝝅 𝒕=𝒎
Let 𝒙 = 𝒎 − 𝒕 ⇒ 𝒅𝒙 = −𝒅𝒕 ⇒ {𝒙 = 𝝅 ⇒ {
𝒎 𝒕=𝟎
𝟎
𝝅 𝝅
𝛀(𝒎) = ∫ 𝐬𝐢𝐧(𝝅 − 𝒎𝒕) ⋅ 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( ) ⋅ 𝐭𝐚𝐧 ( − 𝒕)) (−𝒅𝒕) =
𝝅 𝒎 𝒎
𝒎
𝝅 𝝅
𝒎 𝝅 𝐭𝐚𝐧 (𝒎) − 𝐭𝐚𝐧 𝒕
= ∫ 𝐬𝐢𝐧(𝒎𝒕) ⋅ 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( ) ⋅ ) 𝒅𝒕 =
𝟎 𝒎 𝟏 + 𝐭𝐚𝐧 ( 𝝅 ) ⋅ 𝐭𝐚𝐧 𝒕
𝒎
𝝅 𝝅 𝝅 𝝅 𝝅
𝒎 𝟏 + 𝐭𝐚𝐧 (𝒎) ⋅ 𝐭𝐚𝐧 ( 𝒕 ) + 𝐭𝐚𝐧𝟐 (𝒎) − 𝐭𝐚𝐧 (𝒎) ⋅ 𝐭𝐚𝐧 𝒕
= ∫ 𝐬𝐢𝐧(𝒎𝒕) ⋅ 𝐥𝐨𝐠 ( 𝝅 ) 𝒅𝒕 =
𝟎 𝟏 + 𝐭𝐚𝐧 (𝒎) ⋅ 𝐭𝐚𝐧 𝒕
𝝅 𝟐𝝅
𝒎 𝟏 + 𝐭𝐚𝐧 ( 𝒎 )
= ∫ 𝐬𝐢𝐧(𝒎𝒕) ⋅ 𝐥𝐨𝐠 ( 𝝅 ) 𝒅𝒕 =
𝟎 𝟏 + 𝐭𝐚𝐧 (𝒎) ⋅ 𝐭𝐚𝐧 𝒕
𝝅 𝝅
𝒎 𝟐𝝅 𝒎 𝝅
= ∫ 𝐬𝐢𝐧(𝒎𝒕) ⋅ 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( )) 𝒅𝒕 − ∫ 𝐬𝐢𝐧(𝒎𝒕) ⋅ 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( ) ⋅ 𝐭𝐚𝐧 𝒕)
𝟎 𝒎 𝟎 𝒎
𝝅
𝝅 𝒎
𝛀(𝒎) = 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧𝟐 ( )) ∫ 𝐬𝐢𝐧(𝒎𝒙) 𝒅𝒙 − 𝛀(𝒎)
𝒎 𝟎
Hence,
𝝅
𝝅𝟐
𝟏 𝒎
𝟐𝛀(𝒎) = 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( )) ⋅ (− ) 𝐜𝐨𝐬(𝒎𝒙)| =
𝒎 𝒎 𝟎
𝟏 𝟏 𝝅
= 𝐥𝐨𝐠 ( 𝝅 ) (− ) (𝐜𝐨𝐬 (𝒎 ⋅ ) − 𝐜𝐨𝐬 𝟎) =
𝐜𝐨𝐬 𝟐 (𝒎) 𝒎 𝒎
𝟏 𝝅 𝟏 𝟏 𝝅
− 𝐥𝐨𝐠 (𝐜𝐨𝐬 ( )) ⋅ (− ) (𝐜𝐨𝐬 𝝅 − 𝐜𝐨𝐬 𝟎) = 𝐥𝐨𝐠 (𝐜𝐨𝐬 ( )) ⋅ (−𝟐)
𝟐 𝒎 𝒎 𝟐𝒎 𝒎
Therefore,
𝟏 𝝅
𝛀(𝒎) = − 𝐥𝐨𝐠 (𝐜𝐨𝐬 ( ))
𝟐𝒎 𝒎

www.ssmrmh.ro
𝒂 𝒂
𝐔𝐬𝐢𝐧𝐠: ∫ 𝒇(𝒙) 𝒅𝒙 = ∫ 𝒇(𝒂 − 𝒙) 𝒅𝒙 𝐰𝐞 𝐠𝐞𝐭:
𝟎 𝟎
𝝅
𝒎 𝝅 𝝅 𝝅
𝛀(𝒎) = ∫ 𝐬𝐢𝐧 [𝒎 ( − 𝒙)] ⋅ 𝐥𝐨𝐠 [𝟏 + 𝐭𝐚𝐧 ( ) 𝐭𝐚𝐧 ( − 𝒙)] 𝒅𝒙 =
𝟎 𝒎 𝒎 𝒎
𝝅 𝝅 𝝅
𝒎 𝐭𝐚𝐧 (𝒎) [𝐭𝐚𝐧 (𝒎) − 𝐭𝐚𝐧 𝒙]
= ∫ 𝐬𝐢𝐧(𝒎𝒙) ⋅ 𝐥𝐨𝐠 [𝟏 + 𝝅 ] 𝒅𝒙 =
𝟎 𝟏 + 𝐭𝐚𝐧 (𝒎) ⋅ 𝐭𝐚𝐧 𝒙
𝝅
𝒎 𝝅
= ∫ 𝐬𝐢𝐧(𝒎𝒙) ⋅ 𝐥𝐨𝐠 (𝐬𝐞𝐜 𝟐 ( )) 𝒅𝒙 − 𝛀(𝒎)
𝟎 𝒎
Hence,
𝝅
𝟐 𝝅 𝒎 𝟐 𝝅
𝟐𝛀(𝒎) = 𝐥𝐨𝐠 (𝐬𝐞𝐜 ( )) {− 𝐜𝐨𝐬(𝒎𝒙)}| = 𝐥𝐨𝐠 (𝐬𝐞𝐜 ( )) (𝟏 + 𝟏)
𝒎 𝒎 𝟎 𝒎 𝒎
Therefore,
𝟐 𝝅
𝛀(𝒎) = 𝐥𝐨𝐠 (𝐬𝐞𝐜 ( ))
𝒎 𝒎
Solution 4 by Ankush Kumar Parcha-India
𝝅
𝒎 𝝅
𝛀(𝒎) = ∫ 𝐬𝐢𝐧(𝒎𝒙) ⋅ 𝐥𝐨𝐠 [𝟏 + 𝐭𝐚𝐧 ( ) 𝐭𝐚𝐧 𝒙] 𝒅𝒙 =
𝟎 𝒎
𝝅 𝝅 𝝅
𝟏 𝝅 𝒎 𝐭𝐚𝐧 (𝒎) 𝒎 𝐬𝐞𝐜 𝟐 𝒙 ⋅ 𝐜𝐨𝐬(𝒎𝒙)
= − 𝐜𝐨𝐬(𝒎𝒙) ⋅ 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( ) 𝐭𝐚𝐧 𝒙)| − ∫ 𝝅 𝒅𝒙 =
𝒎 𝒎 𝟎 𝒎 𝟎 𝟏 + 𝐭𝐚𝐧 ( ) 𝐭𝐚𝐧 𝒙
𝟒
𝝅
𝟐 𝝅 𝟏 𝝅 𝟒
= 𝐥𝐨𝐠 (𝐬𝐞𝐜 ( )) − 𝐜𝐨𝐬(𝒎𝒙) ⋅ 𝐥𝐨𝐠 (𝟏 + 𝐭𝐚𝐧 ( ) 𝐭𝐚𝐧 𝒙)| − 𝛀(𝒎)
𝒎 𝒎 𝒎 𝒎 𝟎
Therefore,
𝟒 𝝅 𝟐 𝝅
𝟐𝛀(𝒎) = 𝐥𝐨𝐠 (𝐬𝐞𝐜 ( )) ⇒ 𝛀(𝒎) = 𝐥𝐨𝐠 (𝐬𝐞𝐜 ( ))
𝒎 𝒎 𝒎 𝒎

www.ssmrmh.ro
It’s nice to be important but more important it’s to be nice.

At this paper works a TEAM.
This is RMM TEAM.
To be continued!
Daniel Sitaru

27 RMM Winter Edition 2022 Solutions

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

27 RMM Winter Edition 2022 Solutions

Uploaded by

Copyright:

Available Formats

Number 27 WINTER 2022

1 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

2 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

Proposed by Marian Ursărescu-Romania

3 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

𝑾𝒊𝒕𝒉 𝒆𝒒𝒖𝒂𝒍𝒊𝒕𝒚 𝒊𝒇 𝜟𝑨𝑩𝑪 𝒊𝒔 𝒆𝒒𝒖𝒊𝒍𝒂𝒕𝒆𝒓𝒂𝒍.

4 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

JP.393 In 𝚫𝑨𝑩𝑪 the following relationship holds:

Solution 2 by Mohamed Amine Ben Ajiba-Tanger-Morocco

5 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

𝑭 𝟐(𝒃𝟐 + 𝒄𝟐 ) − 𝒂𝟐 𝑭 𝟐∑𝒃𝒄(𝒃𝟐 + 𝒄𝟐 ) − 𝒂𝒃𝒄∑𝒂 𝟐∑𝒃𝒄(𝒃𝟐 + 𝒄𝟐 ) − 𝒂𝒃𝒄∑𝒂

6 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

Equality holds for 𝒙 = 𝒚 = 𝒛 = 𝒕 = 𝟏.

𝟒 (𝒙𝟒 + 𝟏)(𝒚𝟒 + 𝟏)(𝒛𝟒 + 𝟏)(𝒕𝟒 + 𝟏)

[(𝒙𝟒 + 𝟏)(𝒚𝟒 + 𝟏)(𝒛𝟒 + 𝟏)(𝒕𝟒 + 𝟏)]𝟐 ≥

for 𝒙 ≥ 𝟏: 𝒙𝟔 + 𝟑𝒙𝟐 ≥ 𝟑𝒙𝟒 + 𝟏 and for 𝒙 < 1: 𝒙𝟔 + 𝟑𝒙𝟐 < 3𝒙𝟒 + 𝟏

Solution 3 by Ravi Prakash-New Delhi-India

Solution 4 by Mohamed Amine Ben Ajiba-Tanger-Morocco

Solution 5 by Daniel Văcaru-Romania

8 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

JP.395 Let 𝑹 and 𝒓 be the circumradius and inradius, respectively, of triangle

Proposed by George Apostolopoulos-Messolonghi-Greece

9 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

10 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

Proposed by George Apostolopoulos-Messolonghi-Greece

12 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

Solution 2 by Mohamed Amine Ben Ajiba-Tanger-Morocco

13 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

14 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

Solution 2 by Florentin Vişescu-Romania

⇒ 𝐥𝐨𝐠 𝟐 (𝟑𝒙 − 𝟏) = 𝒇−𝟏 (𝒇−𝟏 (𝒙)) ⇒ 𝒇 (𝒇(𝒇(𝒙))) = 𝒙

Lemma. If 𝒇 is strictly increasing function with 𝒇 (𝒇(𝒇(𝒙))) = 𝒂, then 𝒇(𝒂) = 𝒂.

𝒇 (𝒇(𝒇(𝒂))) > 𝑓 (𝒂) > 𝑎 impossible!

𝒇 (𝒇(𝒇(𝒂))) < 𝑓 (𝒂) < 𝑎 impossible!

Return to the problem. 𝒇 (𝒇(𝒇(𝒙))) = 𝒙 ⇒ 𝒇(𝒙) = 𝒙 ⇒ 𝐥𝐨𝐠 𝟐 (𝟑𝒙 − 𝟏) = 𝒙 ⇒ 𝟑𝒙 = 𝟐𝒙 + 𝟏

Solution 3 by Agayev Seddredin-Azerbaijan

15 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

16 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

JP.398 If 𝒂, 𝒃, 𝒄 > 0; 𝑎𝑏√𝒂𝒃 + 𝒃𝒄√𝒃𝒄 + 𝒄𝒂√𝒄𝒂 = 𝟑 then:

Proposed by Daniel Sitaru-Romania

𝒂𝟑 + 𝒃𝟑 − 𝟐𝒂𝒃√𝒂𝒃 + 𝟖𝒂𝒃√𝒂𝒃 ≤ 𝒂𝟑 + 𝒃𝟑 + 𝟑𝒂𝒃(𝒂 + 𝒃) ⇔

Equality holds for 𝒂 = 𝒃 = 𝒄.

17 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

Equality holds for 𝒂 = 𝒃 = 𝒄.

∑(𝒂 + 𝒃)𝟑 = ∑(𝒂𝟑 + 𝒃𝟑 + 𝟑𝒂𝟐 𝒃 + 𝟑𝒂𝒃𝟐 ) = 𝟐 ∑ 𝒂𝟑 + 𝟑 ∑ 𝒂𝒃(𝒂 + 𝒃) ; (𝟏)

𝟑 ∑ 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟏𝟖 ⇔ ∑ 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟔; (𝟑)

Equality holds for 𝒂 = 𝒃 = 𝒄.

18 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

Equality holds for 𝒂 = 𝒃 = 𝒄.

(𝒂𝒃)𝟐 (𝒃𝒄)𝟐 (𝒄𝒂)𝟐 𝟏

𝑩𝒆𝒓𝒈𝒔𝒕𝒓𝒐𝒎 (𝒂𝒃 + 𝒃𝒄 + 𝒄𝒂)𝟐 𝟏

Solution 2 by Sanong Huayrerai-Nakon Pathom-Thailand

Solution 4 by Avishek Mitra-West Bengal-India

∑ 𝒂𝟐 𝒃𝟐 ≥ 𝟏𝟔𝑭𝟐 (𝒏𝒆𝒆𝒅 𝒕𝒐 𝒑𝒓𝒐𝒗𝒆)

20 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

𝟑(𝒂𝟑 − 𝒃𝟑 ) − 𝟑𝒂𝒃(𝒂 + 𝒃) ≥ 𝟎 ⇔ (𝒂 + 𝒃)(𝒂𝟐 − 𝒂𝒃 + 𝒃𝟐 ) − 𝒂𝒃(𝒂 + 𝒃) ≥ 𝟎

Equality holds for 𝒙 = 𝒚 = 𝒛 = 𝟏.

Equality holds for 𝒙 = 𝒚 = 𝒛 = 𝟏.

By power mean arithmetic mean inequality:

Equality holds for 𝒂 = 𝒃 = 𝒄 = 𝟏 ⇔ 𝒙 = 𝒚 = 𝒛 = 𝟏

22 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

23 NUMBER 26-RMM AUTUMN EDITION 2022-SOLUTIONS

JP.402 In acute 𝚫𝑨𝑩𝑪 the following relationship holds: