Professional Documents
Culture Documents
CHE 536 Engineering Optimization: Course Policies and Outline
CHE 536 Engineering Optimization: Course Policies and Outline
ENGINEERING OPTIMIZATION
COURSE POLICIES AND
OUTLINE
Prof. Shi-Shang Jang
Chemical Engineering Department
National Tsing-Hua University
Hsinchu
Feb. 2011
Instructor: Prof. Shi-Shang Jang
Class Meeting: Every Thursday, 2:00-5:00
Classroom: Rm#221, Chem. Engng. Buldg.
Textbook : P. Venkataraman, Applied Optimization with MATLAB
Programming. Wiley-Interscience, 2001
References:
Reklaitis, Ravindran and Ragsdell, Engineering Optimization, Methods
and Applications, .Wiley, 1983.
Vanderplaats, G.N., Numerical Optimization Techniques for Engineering
Design with Applications, McGraw-Hill International , 1993.
Optimization of Chemical Processes, 2nd Ed., Edgar, Himmelblau and
Lasdon, McGraw-Hill, Boston, 2001
Course Objective:
To learn problem formulation of optimization.
To realize the algorithms of numerical methods of optimization.
To know the applications of numerical optimization.
Policies
Why optimized?
Improve the process to realize
maximum system potential;
Attain improved designs;
maximize profits; reduce cost of
productions.
Essential features of optimization
problem
An objective function is defined which needs to be
either maximized or minimized. The objective function
may be technical or economic. Examples of economic
objective are profits, costs of production etc..
Technical objective may be the yield from the reactor
that needs to be maximized, minimum size of an
equipment etc.. Technical objectives are ultimately
related to economics.
Underdetermined system: If all the design variables
are fixed. There is no optimization. Thus one or more
variables is relaxed and the system becomes an
underdetermined system which has at least in
principle infinite number of solutions.
Essential features of optimization
problem- Continued
Competing influences: In most of the optimization
problems, there would be some set of variables which
has opposite influence on the objective function. Such
competing influences require some balancing and
hence result in typical optimization problems.
Restrictions: Usually the optimization is done keeping
certain restrictions or constraints. Thus, the amount of
row material may be fixed or there may be other
design restrictions. Hence in most problems the
absolute minimum or maximum is not needed but a
restricted optimum i.e. the best possible in the given
condition
Problem Statements
x x1 , x2 ,, xn
T
Given a design vector
An objective function f(x)
A set of equality constraints g(x)=0
A set of inequality constraints, h(x)0
The general problem formulation:
min f x
x
s.t. gx 0
h ( x) 0
Examples- An Inspector Problem
(Linear Programming):
Assume that it is desired to hire some inspectors for
monitoring a production line. A total amount of 1800
species of products are manufactured every day (8
working hours), while two grades of inspectors can be
found. Maximum, 8 grade A inspector and 10 grade B
inspector are available from the job market. Grade A
inspectors can check 25 species/hour, with an
accuracy of 98 percent. Grade B inspectors can
check 15 species/hour, with an accuracy of 95 percent.
The wage of a grade A inspector is $4.00/hour, and
the wage of a grade B inspector is $3.00/hour. What
is the optimum policy for hiring the inspectors?
Problem Formulation
Assume that the x1 grade A inspectors x2 grade B
inspects are hired, then
total cost to be minimized
48 x1 +38 x2 +2580.022 x1 +1580.052
x2
=40 x1 +36 x2
manufacturing constraint
258 x1 +158 x2 1,800 200 x1 +120 x2 1,800
no. of inspectors available:
0 x1 8
0 x2 10
The Graphical Solution
15
X 2=10 B (8,10)
10
C (3,10)
40
X
1 +3
6X
5X 1
40 2 =Z
X
+3
+3 =6
X2
1
6X 00
=4
5 2 =Z
5
Z= =5
38 00
0
X 1=8
(8,5/3)
5 10
Example optimal pipe diameter
cost
1.92e-9*(l*q^2.68)
/(d(i)^4.68);
f(i)=0.45*l+0.245*l*d(
i)^1.5+3.25*hp^0.5+
61.6*hp^0.925+102; diameter
end
Examples- A Chemical Reactor
Design Problem
A B C k1 k10 e E1 / RT
k1 k2
k 2 k 20 e E 2 / RT
Material Balances:
dC A
k1C A
dt
dC B
k1C A k 2 C B
dt
dC C
k 2C B
dt
Energy Balances:
H 1 k1C AV H 2 k 2 C BV UAT TJ
dT
VC P
dt
Examples- A Chemical Reactor
Design Problem - Continued
Problem: Assume that C0, TJ, tf can be
designed, it is our objective to find good
settings such that CB(tf) can be
maximized
Special Types of NLP problems
Unconstrained optimization
min f x In here, neither equality nor inequality constraint
x exists
xi 0
In here, all the objective function and constraints
are linear functions to x.
Special Types of NLP problems -
Continued
Quadratic Programming (QP)
max f x a 0 a1 x x T Qx
x
n
s.t. aij xi b j 0 j 1, , m
i 1
xi 0
is minimized
Special Types of NLP problems -
Continued
Dynamic Programming (DP)
Here, xi represents the state of the system in the ith
stage and ui is some decision variable, i.e. some decision
taken at the ith stage. The values of xi and ui determine uniquely xi+1:
xi 1 f xi , ui , i
Contour plot
2.5
1.5 0.1
0.1
0.2 0.3 0.2
1 0.4
0.5
0.5 0.
6
0.1
0.4
0.7
0.3
0.4
0
0.1
7
0.8 0. 5
0. 3
0.2
-0.5 0.
0.
5 0.
2
0.6
-1 0.4
0.3
0.1 0.2
-1.5 0.1
-2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Terminology - Continued
2.5 2
2 4
2
6
1.5
2
1 4
local 0 2 2
minimum 0.5
-2
-0.5 0
2
-1 2 -2 0
0
-4
-1.5
-6
-2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
local
minimum
Terminology - Continued
x1 1 x2 x1 1 x2
Unimodal function
If the function increases up to a maximum then
decreases for all other values, it is called a
unimodal function
Feasible Regions-
Region of search.
The region bounded by the inequality and equality
constraints.
for i=1:100
for j=1:100
z(i,j)=xx(i,j)^2+yy(i,j)^
2;
end
end
mesh(xx,yy,z)
Contour plot of function z=x^2+y^2
contour(xx,yy,z)
Himmelblaus Function
function y=himmel(x1,x2)
y=(x1*x1+x2-11)^2+(x1+x2*x2-7)^2;
x1=linspace(-5,5,21);
x2=linspace(-5,5,21);
for i=1:21
for j=1:21
zz(i,j)=himmel(x1(i),x2(j));
end
end;
mesh(x1,x2,zz);
3-dimensional Plot
2500
2000
1500
1000
500
0
10
5 10
0 5
0
-5 -5
-10 -10
Contour Plot contour(x1,x2,zz)
-6 -4 -2 0 2 4 6
6
-2
-4
-6