You are on page 1of 10

Verification dan Validation

By: Amanda Dwi Wantira, S.Tr., M.T.


Pemodelan Sistem (TI201419)
Teknik Industri – Institut Teknologi Kalimantan © 2021
Introduction
• After building an influence diagram, a mathematical model is built. Such
model may consist of numerous equations and inequalities. Notice that
the model is not real thing (abstract world).

• The final version of model may be the result of a process of enrichment


and re-formulation.

• To finding the preferred solution, sometimes we need to manipulate the


model in order to explore the response of the system performance to
changes in controllable and uncontrollable inputs.

• We will explore set of feasible solutions or the solution space.


Testing and Validation

Has two facets:


1. Internal validity (Verification)
2. External validity (Validation)
Internal Validity (Verification)
› Checking that the model is logically and mathematically correct
and that the data used is correct.

› Validation aims to determine the suitability of the simulation results


with the symptoms or processes that are imitated.

› A valid model is if its basic structure and pattern can describe the
behavior of the real system, or can represent fairly accurately, the
data collected with respect to the real system or assumptions
made based on references to the way the real system works.
Verification Methods
› Inspection
› Demonstration
› Testing
› Analysis
External Validity (Validation)
› Validation is an attempt to conclude whether the system
model is a legitimate representative of the reality being
studied, so that it can produce convincing conclusions.
Relations
Points of Verification and Validation
Model Verification Validation
Does the model contain all the appropriate
elements, events, and relations?
Conceptual
Can the model answer the modeling question?

Are events represented correctly? Does the model contain all events in the
conceptual model?

Are mathematical formulas and relations


Logic correct?

Are statistical measures formulated correctly? Does the model contain all the relationships that
exist in the conceptual model?

Does computer code render all aspects of logic Does computer mode duplicate real system
model? performance?
Are statistics and formulas calculated Can the computer model represent the validity of
Simulation
correctly? a real system?
Does the mode contain coding errors? Does the computer model output have credibility
with system experts and decision makers?
Points of Verification and Validation
› Verifikasi Satuan
(case menggunakan soal UTS kemarin)

Fungsi tujuan model


Minimasi TC=FC+VC+BPM+BPT+BPP

= ෍ ෍ 𝑧𝑖𝑘 ∗ 𝑓 𝑘 + 𝑅 ෍ ෍ ෍ 𝑡𝑖𝑗
𝑘 𝑘
𝑥𝑖𝑗 + 𝛼 ෍ ෍ 𝑤𝑖𝑘 + 𝛽 ෍ ෍ 𝑑𝑖𝑘 + 𝜎 𝑛 − ෍ ෍ ෍ 𝑥𝑖𝑗
𝑘

𝑖∈𝑉 𝑘∈𝐾 𝑖∈𝑉 𝑗∈𝑉 𝑘∈𝐾 𝑖∈𝑉 𝑘∈𝐾 𝑖∈𝑉 𝑘∈𝐾 𝑖∈𝑉 𝑗∈𝑉 𝑘∈𝐾
𝑖=0 𝑖≠𝑗 𝑖≠𝑗

Dimana:
› 𝑧𝑖𝑘 Binary untuk kendaraan yang digunakan
› 𝑓 𝑘 Biaya tetap yang dikeluarkan untuk menggunakan
kendaraan k (rupiah/unit)
› 𝑅 Biaya perjalan per unit waktu (rupiah/menit)
› 𝑘
𝑡𝑖𝑗 Waktu tempuh antara lokasi i ke lokasi j (menit)
menggunakan kendaraan k
› 𝑘
𝑥𝑖𝑗 Binary jika titik rute dilalui oleh kendaraan k dan 0 jika
lainnya
Points of Verification and Validation
Komponen
• Biaya tetap: biaya yang dipengaruhi oleh jumlah kendaraan dan • Biaya penalti keterlambatan: biaya yang dipengaruhi total
jenis kendaraan yang digunakan waktu kendaraan terlambat tiba di pelanggan

𝐹𝐶 = ෍ ෍ 𝑧𝑖𝑘 ∗ 𝑓 𝑘 𝑅𝑢𝑝𝑖𝑎ℎ 𝐹𝐶 = 𝛽 ෍ ෍ 𝑑𝑖𝑘 𝑅𝑢𝑝𝑖𝑎ℎ


𝑅𝑢𝑝𝑖𝑎ℎ = 𝑈𝑛𝑖𝑡 × … 𝑠𝑒𝑠𝑢𝑎𝑖 𝑅𝑢𝑝𝑖𝑎ℎ = × 𝑀𝑒𝑛𝑖𝑡 … 𝑠𝑒𝑠𝑢𝑎𝑖
𝑖∈𝑉 𝑘∈𝐾 𝑈𝑛𝑖𝑡 𝑖∈𝑉 𝑘∈𝐾 𝑀𝑒𝑛𝑖𝑡
𝑖=0

• Biaya variabel: biaya yang dipengaruhi oleh total durasi tur • Biaya penalti untuk pelanggan yang tidak terlayani karena
pelanggaran batasan kapasitas dan jendela waktu
𝐹𝐶 = 𝑅 ෍ ෍ ෍ 𝑡𝑖𝑗𝑘 𝑥𝑖𝑗
𝑘 𝑅𝑢𝑝𝑖𝑎ℎ
𝑅𝑢𝑝𝑖𝑎ℎ = × 𝑀𝑒𝑛𝑖𝑡 × 𝑈𝑛𝑖𝑡𝑙𝑒𝑠𝑠 … 𝑠𝑒𝑠𝑢𝑎𝑖
𝑖∈𝑉 𝑗∈𝑉 𝑘∈𝐾 𝑀𝑒𝑛𝑖𝑡 𝑘
𝑖≠𝑗 𝐹𝐶 = 𝜎 𝑛 − ෍ ෍ ෍ 𝑥𝑖𝑗
𝑖∈𝑉 𝑗∈𝑉 𝑘∈𝐾
𝑖≠𝑗
• Biaya penalti menunggu: biaya yang dipengaruhi oleh total
waktu kendaraan menunggu di pelanggan 𝑅𝑢𝑝𝑖𝑎ℎ = 𝑅𝑢𝑝𝑖𝑎ℎ × 𝑢𝑛𝑖𝑡𝑙𝑒𝑠 − 𝑢𝑛𝑖𝑡𝑙𝑒𝑠𝑠 … 𝑠𝑒𝑠𝑢𝑎𝑖

𝑅𝑢𝑝𝑖𝑎ℎ
𝐵𝑃𝑀 = 𝛼 ෍ ෍ 𝑤𝑖𝑘 𝑅𝑢𝑝𝑖𝑎ℎ = × 𝑀𝑒𝑛𝑖𝑡 … 𝑠𝑒𝑠𝑢𝑎𝑖
𝑀𝑒𝑛𝑖𝑡
𝑖∈𝑉 𝑘∈𝐾

You might also like