You are on page 1of 20

 

Visvesvaraya National Institute of Technology, Nagpur


Department of Chemical Engineering
First Sessional Examination – 2012
Subject: Optimization Techanique
Maximum Marks: 15 Time: 1 Hour

Q1a. Bracket the maximum of function 10 2 5 using bounding phase method with initial guess
1.25 and ∆ 0.3. Summarize all final answers in tabulated form. (3)

Q1b. Maximize the function 10 2 5 using Fibonacci series method up to 0.9 accuracy in x with
bracket calculated in Q1a. Summarize all final answers in tabulated form. (5)

Q2. Optimize the function ⁄log using Newton’s method with initial guess 2, do the calculation at least for
three iterations, use ∆ 0.01 for calculating gradient. Summarize all final answers in tabulated form. (5)

Q2. Discuss the necessary and sufficient condition for optimization. (2)

Visvesvaraya National Institute of Technology, Nagpur


Department of Chemical Engineering
First Sessional Re-Examination – 2012
Subject: Optimization Technique
Maximum Marks: 15 Time: 1 Hour

Q1. Considers minimizes the function 1.5 bracket the solution using extensive search method starting at -2
and spacing interval length is 0.5 use this bracket and optimize to within ±0.05 by (a) golden section (b) Fibonacci method
(9)

Q2. Considers minimizes the function 5 20 5

(a) Bracket the solution using extensive search method starting at 0 and spacing interval length is 0.1
(b) Optimize the given problem using interval halving method in the interval (0, 5) (6)

Visvesvaraya National Institute of Technology, Nagpur


Department of Chemical Engineering
Second Sessional Examination – 2012
Subject: Optimization Techanique
Maximum Marks: 15 Time: 1 Hour

Q1. What are vertices of hypercube with side length 1 unit and one point at origin for 1D, 2D, 3D, hypercube? (2)

Q2. Explain how Single Variable optimization techniques can be used for multivariable problems? (2)

Q3. Explain in detail with figure, examples and algorithm Powell’s Conjugate direction method for multivariable
optimization (5)


 
 
Q4. Find the minimum of function , 2 2 using simplex method with initial simplex
of three points as 0,0 , 2,0 , 1,1 with expansion factor 1.5 and contraction factor 0.5. Show
the detail calculations of at least two iterations. (6)

Visvesvaraya National Institute of Technology, Nagpur


Department of Chemical Engineering
Final Examination – April 2012
Subject: Optimization Techniques
Maximum Marks: 60 Time: 3 Hour

Note: Give proper explanation wherever needed


Report the results in proper table from and highlight important results wherever needed

Q1. Maximize f(x) = 3x1+4.67x2 using graphical method


Subjected to constrains 10x1+5x2 ≤ 2500, 4x1+10x2 ≤ 2000, 5x1+6x2 ≤ 1500, x1,x2 ≥ 0
(5)
Q2. Minimize f(x) = 3x1+4x2 using tabulated form simplex method.
Subjected to constrains 2x1+3x2 ≥ 8, 5x1+2x2 ≥ 12, x1,x2 ≥ 0 (5)

Q3. Maximize f(x) = 6x1+5x2 using algebraic form simplex method.


Subjected to constrains x1+x2 ≤ 5, 3x1+2x2 ≤ 12, x1,x2 ≥ 0 (5)

Q4. Minimize f(x) = 100 1 with staring point [-1,1]T along the direction [4,0]T
a) Bracket the optimum solution using bounding phase method with Δ=0.2 (5)
b) Using bracket obtained in part (a) find the optimum point using interval halving method for 3 iterations.
(5)
Q5. Minimizes 54/ using secant method with initial interval 1 and 7 for at least for 3 iterations.
(5)
Q6. Compares the golden section search method and Fibonacci search method for 5 iteration for minimization of
10exp 0.1 in the interval (-10, 5) (10)

Q7. Minimize 11 7 using Hooke-Jeeves pattern search method with initial point at
[0, 0]T and increment vector is [0.5, 0.5]T and reduction factor is 2. Continue the calculation till increment vector reduce to
value [0.125, 0.125]T. (10)

Q8. Minimize 3 6 by evolutionary optimization method, starting point at [2,-1,1]T with size of
hypercube [1,1,1] with reduction factor 0.5. Do the calculation for 3 iterations (10)


 
 
Visvesvaraya National Institute of Technology, Nagpur
Department of Chemical Engineering
First Sessional Examination – 2013
Subject: Optimization Techanique
Maximum Marks: 15 Time: 1 Hour

Q1. Use three iterations of golden section search method in order to maximize the function in the interval (-5, 5)
10 2 5 (4)
Q2. Bracket the minimum of the following function using the bounding phase method with an initial guess and
initial interval ∆=1, 400 10 (4)
Q3. Use three iteration of the bisection and secant method, to minimize the following function in the interval (-10, 5).
10 0.1
Compare the algorithm in terms of the interval and optimal solution obtained at end of every iteration. (7)

Visvesvaraya National Institute of Technology, Nagpur


Department of Chemical Engineering
Second Sessional Examination – 2013
Subject: Optimization Techanique
Maximum Marks: 15 Time: 1 Hour

Q1. Consider 1
a) Discuss in detail the necessary and sufficient condition for optimization for the function given above and note the
optimal point. (4)
b) Compare and discuss couchy’s and Newton’s search direction for same given function towards optimal point from
point [5, 5]T and from point [1.2, 1.2]T. (4)
Q2. Explain in brief Box evolutionary optimization method. Minimize the function using this
T T
method for two iteration’s with initial guess [2, 2, 2, 2] , hypercube size [4, 4, 4, 4] , and size reduction parameter [0.5,
0.5, 0.5, 0.5]T. Report all the result in tabular form. (7)


 
 
Visvesvaraya National Institute of Technology, Nagpur
Department of Chemical Engineering
Final Examination – April 2013
Subject: Optimization Techniques
Maximum Marks: 60 Time: 3 Hour

Note: Give proper explanation wherever needed


Report result in proper table from and highlight important results wherever needed

Q1. Minimize and maximize 2 5 using graphical method with following constrains equations. [5]
6, 0, 8, 18, 12 8, 2, 10

Q2. Explain with algorithm, interval halving method for optimization. Minimize the function for three
iterations in interval [-2, 5] using interval halving method, also give the accuracy value. [5]

Q3. Minimize f(x) = 100 1 with staring point [5, 7]T along the direction [2,3]T
a) Bracket the optimum solution using exhaustive search method using Δ=0.6 [5]
b) Improve on accuracy obtained in part (a) up to 0.3 using Fibonacci search method. [5]

Q4. Minimizes 54/ using successive quadratic estimation method with initial guess 1 and Δ=1, till
variable accuracy of 0.05. [5]

Q5. Explain in details with algorithm simplex search method and Hooke-Jeeves pattern search method for multivariable
nonlinear optimization. [5]

Q6. Explain Marquardt’s method in details with algorithm. Minimize function using Marquardt’s method with initial
guess [6, 4]T and initial λ value 100 for 3 iterations, 11 7 [10]

Q7. Categorized and explain the terms degeneracy, alternate optimum, unboundness and infeasibility among the following
given sets of problem using only simplex method [20]
a) Minimize 4 Subjected to constrain 6, 10, , 0
b) Minimize 4 4 Subjected to constrain 6, 10, , 0
c) Minimize 4 Subjected to constrain 11 , 10, , 0
d) Minimize 4 Subjected to constrain 6, 10, , 0
e) Minimize 4 Subjected to constrain 6, 6, , 0


 
 
Visvesvaraya National Institute of Technology, Nagpur
Department of Chemical Engineering
First Sessional Examination – 2015
Subject: Optimization Technique
Maximum Marks: 15 Time: 1 Hour

Discuss the following all question considering minimizes the function 3 Report all results in tabular form
1. Bracket the solution using extensive search method starting at 0 and spacing interval length is 0.1. (2)
2. Bracket the solution using Bounded phase method starting at 0 and spacing interval length is 0.1. (2)
3. Compare method’s used in 1 & 2 with respect to accuracy and actual number of function calculations in 1 & 2. (1)
4. Perform three iteration of Interval Halving Method in the bracket (-5, 5). (2)
5. Perform three iteration of Fibonacci Search Method in the bracket (-5, 5). (3)
6. Perform three iteration of Golden Section Search Method in the bracket (-5, 5). (3)
7. Compare method’s used in 4, 5 & 6 with respect to percentage of region eliminated per actual number of function
calculations in 4, 5 & 6. (2)

Visvesvaraya National Institute of Technology, Nagpur


Department of Chemical Engineering
Second Sessional Examination – 2015
Subject: Optimization Technique
Maximum Marks: 15 Time: 1 Hour

8. Minimize the Himmelblau function , 11 7 using


a) Box Evolutionary Optimization method with initial guess (4.5, 4.5) and hypercube size, ∆=(2,2)T for three
T

iterations. (5)
T T T
b) Simplex search Method with initial guess (4.5, 4.5) ,(3.5, 4) ,(4.5, 3.5) and simplex expansion factor, γ=1.5 and
simplex contraction factor, β= 0.5 for three iteration (5)
9. Explain briefly with algorithm and compare Cauchy’s method, Newton’s method and Marquardt’s method for
nonlinear multivariable unconstrained optimization. (5)


 
 
Visvesvaraya National Institute of Technology, Nagpur
Department of Chemical Engineering
Final Examination – April 2015
Subject: Optimization Techniques
Maximum Marks: 60 Time: 3 Hour

Note: Give proper explanation wherever needed


Report result in proper table from and highlight important results wherever needed

Note: Use only Simplex method to solve Q1. – Q4


Q1. Minimize 5 4 Subjected to constraints 2 3, 9 4 9, , 0 [3]
Q2. Maximize 5 4 Subjected to constraints 2 3, 9 4 9, , 0 [3]
Q3. Minimize 5 4 Subjected to constraints 2 3, 9 4 9,
7 4 57, , 0 [5]
Q4. Maximize 5 4 Subjected to constraints 2 3, 9 4 9,
7 4 57, , 0 [5]
Q5. Minimize 1 2 Subjected to constraitn 4 4 1
Using Penalty Function Method with penalty term of parabolic penalty type
If needed use of analytical differentiation is permitted to solve this problem
1). Iteration one Penalty Function Method
a). Use Cauchy’s direction with initial guess (2, 5) and penalty parameter 0.1 [2]
b). Bracket the minimum using Bounded Phase Method with Δ=0.5 [3]
c). Improve the accuracy using Region Halving Method for two iteration [3]
2). Iteration two Penalty Function Method
a). Use Newton’s direction with improvise value of penalty parameter with increment or reduction factor
of 10. [2]
b). Bracket the minimum using Extensive Search Method with Δ=0.5 [3]
c). Improve the accuracy using Golden Section Method for two iteration [3]
3). Iteration three Penalty Function Method
a). Use Marqurdt’s direction with λ=25 and improvise value of penalty parameter with increment or
reduction factor of 10. [3]


 
 
Visvesvaraya National Institute of Technology, Nagpur

Department of Chemical Engineering

First Sessional Examination – 2015


Subject: Optimization Techniques in Process Design (CML-529)

Maximum Marks: 15 Time: 1 Hour

Explain with basic principle, write algorithm and solve given problem for all questions below (5 X 3)
.
Q1) Bounding phase method . Given Problem: Minimize function 3 0.01 Bracket
optimum diameter using Bounding phase method with initial guess 0.2 m and step size 0.1 m

Q2) Fibonacci search method. Given Problem: Optimize the cost function for two iteration form the results obtained in
(Q1) using Fibonacci search method

Q3) Secant method. Given Problem: Optimize the cost function for two iteration form the results obtained in
(Q2) using Secant method (Numerical or Analytical differentiation is permitted)

_______________________________________________________________________________________

Visvesvaraya National Institute of Technology, Nagpur


Department of Chemical Engineering
Second Sessional Examination: 10-October- 2015
Subject: Optimization Techniques in Process Design (CML-529)
Maximum Marks: 15 Time: 1 Hour

Q1. What is Kuhn Tucker Conditions? How many, total Kuhn Tucker Conditions are for following problem? Check all
Kuhn Tucker Conditions for points [2, 4]T and [3.5 3.5]T for following problem.

Minimize 1 2

Subjected to constraint 4 4 1, 2 4, 2 4, (6)

Q2. Minimize 3 6 by evolutionary optimization method, starting point at [1, 2, 3]T with size of
hypercube [1, 2, 3] with reduction factor 0.5. Do the calculation for 2 iterations . (6)

Q3. List out all the methods used for multivariable optimization. Explain how single variable methods can be used for
optimizing multivariable problem. (3)


 
 
Visvesvaraya National Institute of Technology, Nagpur
M. Tech. Chemical Engineering
Final Examination – November 2015
Subject: Optimization Techniques in Process Design (CML-529)
Maximum Marks: 50 Time: 3 Hour

Note: Give proper explanation wherever needed


Report result in proper table from and highlight important results wherever needed

Q1. Categorized and explain the terms degeneracy, alternate optimum, unboundness and infeasibility among the
following given sets of problem using only simplex method. Also give the graphical interpretation of all terms
using corresponding problem. For all examples consider , 0 [4x5]
f) Minimize 4 3 Subjected to constrain 4 3, 3 12,
g) Minimize 3 4 Subjected to constrain 2 3 8, 5 2 12,
h) Minimize 4 3 Subjected to constrain 2 3 8, 3 2 12,
i) Minimize 4 3 Subjected to constrain 6 5, 3 11,
j) Minimize 4 3 Subjected to constrain 8 6 25, 3 4 15,

Q2. Explain Interval Halving Method (IHM), Fibonacci Search Method (FSM) and Golden Section Method
(GSM). Show how performance of these method is in order IHM<FSM<GSM [4]

Q3. Explain Cauchy’s Steepest Descent Method (CSDM), Newton’s Method (NM) and Marquardt’s Method
(MM) for multivariable optimization with their advantages and disadvantages. [4]

Q4. Explain and compare Penalty Function Method (PFM) and Method of Multiplier (MoM) for non linear
constrain optimization. [4]

Q5. Explain following terms in Genetic Algorithm and its application in optimization. [1x8]
a) Inspiration from Nature for Genetic Algorithm
b) Origin of Genetic Algorithm for technical optimization
c) Coding and decoding of variables
d) Population and generation
e) Fitness function
f) Roulette-Wheel Principle
g) Crossover
h) Mutation

 
 
Q6. Annual cost of pipe line transporting the fluid is given by C = 57 D1.3 + 734D-4.8 where C is
cost in Rs and D is diameter in meter. Find optimum diameter using genetic algorithm in range 0.09 to 5.20
with accuracy of two decimal with four candidates in generation. Strictly use random number table given below
to generate, compulsory single point crossover and mutation (not more than 5 %). Replace the worst of
population with replica of the best to obtained mating pool. Explain in detail with correct logic how this random
number table is used in calculation; highlight all the important operation clearly. Perform two iterations,
compare and discuss the fitness of all generation population. Give your comment on progress of genetic
algorithm for this problem.
[10]


 
 
Visvesvaraya National Institute of Technology, Nagpur
Department of Chemical Engineering
First Sessional Examination –August 31, 2016
Subject: Optimization Techniques in Process Design (CML-529)
Maximum Marks: 15 Time: 1 Hour

Explain with basic principle, write algorithm and solve given problem for all questions below (5 X 3)
.
Q1) Exhaustive search method given problem: Minimize 3 0.01 Bracket optimum point
using Exhaustive search method with initial guess 104 and step size 16.

Q2) Interval halving method given problem: Optimize the function for three iteration form the results obtained in
(Q1) using Interval halving method.

Q3) Successive quadratic estimation method given problem: Optimize the function for two iteration form the
results obtained in (Q2) using Successive quadratic estimation method.

_______________________________________________________________________________________

Visvesvaraya National Institute of Technology, Nagpur


Department of Chemical Engineering
Second Sessional Examination –October 5, 2016
Subject: Optimization Techniques in Process Design (CML-529)
Maximum Marks: 15 Time: 1 Hour

Assume suitable data wherever needed

Q1. Given function , 10 10


a) Discuss in detail the necessary and sufficient condition for optimization for the function given above and note the
optimal point. (2)
b) Compare and discuss couchy’s and Newton’s search direction for same given function towards optimal point from
point [1, 1]T and from point [10.5, 9.5]T. (3)

Q2. Explain in detail with figure, Simplex Method for multivariable optimization
Find the optimum of function , 10 10 using simplex method with initial simplex of three
points as 0,0 , 2,0 , 1,1 with expansion factor 1.5 and contraction factor 0.5. Show the
detail calculation for one iteration (5)

Q3. Explain in detail with figure, Box Evolutionary Optimization method for multivariable optimization
Find the optimum of function , 10 10 using Box Evolutionary Optimization method with
T T
initial guess (9.5, 9.5) and hypercube size, ∆=(2,2) for two iterations. (5)

10 
 
 
Visvesvaraya National Institute of Technology, Nagpur
M. Tech. Chemical Engineering
Final Examination – November 2016
Subject: Optimization Techniques in Process Design (CML-529)
Maximum Marks: 60 Time: 3 Hour

Note: Give proper explanation wherever needed


Report result in proper table from and highlight important results wherever needed

Q1. A refinery distills crude petroleum from two sources, Saudi Arabia and Venezuela, into three main
products: gasoline, jet fuel and lubricants. The two crudes differ in chemical composition and thus yield
different product mixes the remaining 10% of each barrel is lost to refining.

a) Formulate the linear programming optimization problem using data given in figure. [2]
b) Solve the problem using graphical method [5]
c) Solve the problem using simplex method [8]
d) Compare the graphical method and simplex method using results and observation in (a), (b) and (c) [3]
e) Does slack variable appearing in the final optimal solution? If yes give the physical significance of its values
with respect to given problem above. [2]

Q2. Explain the following terms using schematic graphical representation and also explain how simplex method
will behave in these cases [4X2]
a) Degeneracy
b) Alternate optimum
c) Unboundness
d) Infeasibility
11 
 
 

Q3. Solve set of simultaneous equations, 14 and 4 using simplex method [4]

Q4. Minimize using genetic algorithm in range 1.00 to 5.00 with accuracy of two decimal with

four candidates in generation. Strictly use random number table given below to generate, single point crossover
and mutation (not more than 5 %). Replace the worst of population with replica of the best to obtained mating
pool. Explain in detail with correct logic how this random number table is used in calculation; highlight all the
important operation clearly. Perform two iterations, compare and discuss the fitness of all generation
population. Give your comment on progress of genetic algorithm for this problem. [10]

Q5. Minimize by evolutionary optimization method, starting point at [1, 1, -1]T with size
of hypercube [2, 2, 2] with reduction factor 0.5. Do the calculation till size of hypercube reached < 0.25 [10]

Q6. Minimize f(x) = 100 1 with staring point [2, 2]T along the direction [1, 1]T
a) Bracket the optimum solution using bounding phase method with Δ=0.2 [4]
b) Using bracket obtained in part (a) find the optimum point using golden section search method for 3 iterations.
[4]

12 
 
Visvesvaraya National Institute of Technology, Nagpur
Department of Chemical Engineering
First Sessional Examination – 2017
Subject: Optimization Techanique (CML-423)
Maximum Marks: 15 Time: 1 Hour

All questions carry equal marks [5 X 3]


Minimize 11 7 in direction [8, 9]T from the point [11, 11]T
Q1. Bracket the optimum using extensive search method with 0.3 value of ∆.
Q2. Starting from the result of Q1, eliminate the unwanted region for two iteration using interval
halving method.
Q3. Starting from the result of Q2, use successive quadratic estimation method for two iterations.
Visvesvaraya National Institute of Technology, Nagpur
Department of Chemical Engineering
Second Sessional Examination – 2017
Subject: Optimization Techanique (CML-423)
Maximum Marks: 15 Time: 1 Hour

All questions carry equal marks

Q1. List out all the methods used for multivariable optimization. Explain how single variable methods can
be used for optimizing multivariable problem. (3)

Q2. Minimize 3 6 by evolutionary optimization method, starting point at [3, 4, 5]T with
size of hypercube [2 2 2] with reduction factor 0.5. Do the calculation for 2 iterations . (6)

Q3. Find the minimum of function , 2 2 using simplex method with


initial simplex of three points as 0,0 , 2,0 , 1,1 with expansion factor 1.5 and
contraction factor 0.5. Show the detail calculations of at least two iterations. (6)
Visvesvaraya National Institute of Technology, Nagpur
Department of Chemical Engineering
End Sem Examination – 2017
Subject: Optimization Techanique (CML-423)
Maximum Marks: 60 Time: 3 Hour

Q1. Explain the following methods for multivariable optimization in details with their algorithms
a) Hooke-Jeeves Pattern Search Method (3)
b) Powell’s Conjugate Direction method (3)

Q2. Considers minimizes the function 5 20 5


a) Bracket the solution using extensive search method starting at 0 and spacing interval length is 0.1 (3)
b) Optimize the given problem for two iteration using interval halving method in the interval (0, 5) (3)

Q3. Considers minimizes the function 1.5 bracket the solution using extensive search method
starting at -2 and spacing interval length is 0.5 use this bracket and optimize to within ±0.05 by
a) Golden section (4)
b) Fibonacci method (4)

Q4. Minimizes 54/ using successive quadratic estimation method with initial guess 1 and
Δ=1, till variable accuracy of 0.05. (4)

Q5. Consider 1
a) Discuss in detail the necessary and sufficient condition for optimization for the function given above and
note the optimal point. (4)
b) Compare and discuss couchy’s and Newton’s search direction for same given function towards optimal
point from point [5, 5]T and from point [1.2, 1.2]T. (4)

Q6. Minimize and maximize 2 5 using graphical method with following constraints
equations.
6, 0, 8, 18, 12, 8, 2, 10 (4)

Note: Use only Simplex method to solve all questions of Q7


Q7a. Minimize 5 4 Subjected to constraints
2 3, 9 4 9, , 0 (3)

Q7b. Maximize 5 4 Subjected to constraints


2 3, 9 4 9, , 0 (3)
Q7c. Minimize 5 4 Subjected to constraints
2 3, 9 4 9, 7 4 57, , 0 (4)

Q7d. Maximize 5 4 Subjected to constraints


2 3, 9 4 9, 7 4 57, , 0 (4)

8. Minimize 5 10 using genetic algorithm in range 0.00 to 5.11 with accuracy of two
decimal with four candidates in generation. Strictly use random number table given below to generate,
single point crossover and mutation (not more than 5 %). Replace the worst of population with replica of the
best to obtained mating pool. Explain in detail with correct logic how this random number table is used in
calculation; highlight all the important operation clearly. Perform two iterations, compare and discuss the
fitness of all generation population. Give your comment on progress of genetic algorithm for this problem.
(10)
Visvesvaraya National Institute of Technology, Nagpur
Department of Chemical Engineering
First Sessional Examination –August 2018
Subject: Optimization Techniques in Process Design (CML-529)
Maximum Marks: 15 Time: 1 Hour

Q1. Considers optimization of the function 𝑓(𝑥1 , 𝑥2 , 𝑥3 ) = 𝑥12 + 𝑥22 + 𝑥32

a. Discuss extensive search method with algorithm. Bracket above function from point [4, 5, 6]T in
direction [1 , 2, 3]T using Δ value 1. [5]
b. Discuss region halving method with algorithm. Perform two iteration of region halving method from
the bracket obtained in Q1a for above function. [5]
c. Discuss succession quadratic estimations method with algorithm. Perform one full iteration of
succession quadratic estimations method from the data obtained in Q1b for above function and report
the final answer. [5]

Visvesvaraya National Institute of Technology, Nagpur


Department of Chemical Engineering
Second Sessional Examination –October 2018
Subject: Optimization Techniques in Process Design (CML-529)
Maximum Marks: 15 Time: 1 Hour

Q1. What are vertices of hypercube with side length 1 unit and one point at origin for 1D, 2D, 3D,
hypercube? (2)

Q2. List out all the methods used for multivariable optimization. Explain how single variable methods can
be used for optimizing multivariable problem. (2)

Q3. What is Kuhn Tucker Conditions? What is importance of Kuhn Tucker Conditions? How many, total
Kuhn Tucker Conditions are for following problem
Minimize 𝑓 𝑥 = 𝑥1 − 1 2 + 𝑥2 − 2 2
Subjected to constraint 𝑥1 − 4 2 + 𝑥2 − 4 2 ≤ 1, 2 ≤ 𝑥1 ≤ 4, 2 ≤ 𝑥2 ≤ 4, (2)

Q4. Consider 𝑓 = (𝑥2 − 𝑥12 )2 + (1 − 𝑥1 )2


Compare and discuss couchy’s and Newton’s search direction for same given function towards optimal point
from point [5, 5]T and from point [1.2, 1.2]T. (4)

Q5. Minimize the Himmelblau function 𝑓 𝑥1 , 𝑥2 = (𝑥12 + 𝑥2 − 11)2 + (𝑥1 + 𝑥22 − 7)2 using
Simplex search Method with initial guess (4.5, 4.5)T,(3.5, 4)T,(4.5, 3.5)T and simplex expansion factor, γ=1.5
and simplex contraction factor, β= 0.5 for three iteration (5)
Visvesvaraya National Institute of Technology, Nagpur
M. Tech. Chemical Engineering
Final Examination – November 2018
Subject: Optimization Techniques in Process Design (CML-529)
Maximum Marks: 50 Time: 3 Hour
Note: Give proper explanation wherever needed
Report result in proper table from and highlight important results wherever needed

Q1. A refinery distills crude petroleum from two sources, Saudi Arabia and Venezuela, into three main
products: gasoline, jet fuel and lubricants. The two crudes differ in chemical composition and thus yield
different product mixes the remaining 10% of each barrel is lost to refining.

a) Formulate the linear programming optimization problem using data given in figure. [2]
b) Solve the problem using graphical method [4]
c) Solve the problem using simple
x method [8]
d) Compare the graphical method and simplex method using results and observation in (a), (b) and (c) [3]
e) Does slack variable appearing in the final optimal solution? If yes give the physical significance of its values
with respect to given problem above. [2]

2 2 2 2
Q2. Minimize 𝑓 𝑥 = 𝑥1 − 1 + 𝑥2 − 2 Subjected to constraint 𝑥1 − 4 + 𝑥2 − 4 ≤1

Using Penalty Function Method with penalty term of parabolic penalty type

If needed use of analytical differentiation is permitted to solve this problem

1). Iteration one Penalty Function Method


a). Use Cauchy’s direction with initial guess (2, 5) and penalty parameter 0.1 [2]

b). Bracket the minimum using Bounded Phase Method with Δ=0.5 [3]

c). Improve the accuracy using Region Halving Method for two iteration [3]

2). Iteration two Penalty Function Method

a). Use Newton’s direction with improvise value of penalty parameter with increment or
reduction factor of 10. [2]

b). Bracket the minimum using Extensive Search Method with Δ=0.5 [3]

c). Improve the accuracy using Golden Section Method for two iteration [3]

3). Iteration three Penalty Function Method

a). Use Marqurdt’s direction with λ=25 and improvise value of penalty parameter with increment
or reduction factor of 10. [3]

4). Give proper graphical representation of your all answers in all iteration on graph paper with exact
scale. Predict the answer using graphical representation, what is error in your final answer penalty
function method and graphically predicted answer. [2]
Q3. Maximum of function 𝑓 𝑥 = 10 + 𝑥 3 − 2𝑥 − 5𝑒 𝑥 using genetic algorithm in range -2.40 to 0.50 with
accuracy of two decimal with four candidates in generation. Strictly use random number table given below to
generate, single point crossover and mutation (not more than 5 %). Replace the worst of population with replica
of the best to obtained mating pool. Explain in detail with correct logic how this random number table is used in
calculation; highlight all the important operation clearly. Perform two iterations, compare and discuss the fitness
of all generation population. Give your comment on progress of genetic algorithm for this problem. [10]
Visvesvaraya National Institute of Technology, Nagpur
Department of Chemical Engineering
First Sessional Examination –August 2019
Subject: Optimization Techniques in Process Design (CML-529)
Maximum Marks: 15 Time: 1 Hour

Q1. Consider isothermal CFSTR with series reaction as shown in figure 1. The goal is to
maximize CB in the effluent at steady state subjected only the flow rate of feed F can be adjusted. If given
data are 𝑇 𝑖𝑛 = 33 °C, 𝐶𝐴𝑖𝑛 = 100 𝑚𝑜𝑙𝑒 𝑚3 , 𝑇𝐶𝑖𝑛 = 20 °C, 𝑉 = 1.1 𝑚3 , 𝐹𝐶 = 2.2 𝑚3 ℎ, 𝑇𝐶𝑜𝑢𝑡 = 45 °C,
𝑇 = 45 °C, 𝑘1 = 3 𝑠 (−1) , 𝑘2 = 2 𝑠 (−1) [5 X 3]

330 𝐹
a) Formulate the above optimization problem and show 𝐶𝐵 =
𝐹+3.3 𝐹+2.2

b) Starting with initial guess 𝐹 = 0 and incremental value ∆𝐹 = 0.1 bracket the maximum value of 𝐶𝐵 using
bounding phase method.
c) Improve the accuracy of results obtained in Q1b by using Fibonacci search method for three iterations.
d) Starting with initial guess 𝐹 = 2.0 and incremental value ∆𝐹 = 0.1 what will maximum value of 𝐶𝐵 using
successive quadratic estimation method for two iterations.
e) Starting with initial guess 𝐹 = 2.0 what will maximum value of 𝐶𝐵 using Newton-Raphson method for
two iterations. For numerical differentiation use the incremental value ∆𝐹 = 0.1

Figure 1

You might also like