Engineering Optimization: Theory and Practice, Fourth Edition

(Martin Jones) #1
Problems 487

Starting point for Unconstrained
Value of minimization of minimum of
k rk φ (X, rk) φ (X, rk)=X∗k f (X∗k)=fk∗
1 1 (−0.4597,−5.0) (0.2361,−0.5) 0.1295
2 10 (0.2361,−0.5) (0.8322,−0.05) 2.0001


Estimate the optimum solution,X∗andf∗, using a suitable extrapolation technique.

7.35 The results obtained in an exterior penalty function method of solution for the optimization
problem stated in Problem 7.15 are given below:


r 1 = 0. 01 , X∗ 1 =

{
− 0. 80975
− 50. 0

}
, φ 1 ∗= − 24. 9650 , f 1 ∗= − 49. 9977

r 2 = 1. 0 , X∗ 2 =

{
0. 23607
− 0. 5

}
, φ∗ 2 = 0. 9631 , f 2 ∗= 0. 1295

Estimate the optimum design vector and optimum objective function using an extrapola-
tion method.

7.36 The following results have been obtained during an exterior penalty function approach:


r 1 = 10 −^10 , X∗ 1 =

{
0. 66
28. 6

}

r 2 = 10 −^9 , X∗ 2 =

{
1. 57
18. 7

}

Find the optimum solution,X∗, using an extrapolation technique.

7.37 The results obtained in a sequential unconstrained minimization technique (using an exte-
rior penalty function approach) from the starting pointX 1 =


{ 6. 0
30. 0

}
are

r 1 = 10 −^10 , X∗ 1 =

{
0. 66
28. 6

}
; r 2 = 10 −^9 , X∗ 2 =

{
1. 57
18. 7

}

r 3 = 10 −^8 , X∗ 3 =

{
1. 86
18. 8

}

Estimate the optimum solution using a suitable extrapolation technique.

7.38 The two-bar truss shown in Fig. 7.26 is acted on by a varying load whose magnitude
is given byP (θ )=P 0 cos 2θ; 0 ◦≤θ≤ 360 ◦. The bars have a tubular section with
mean diameterdand wall thicknesst. UsingP 0 = 50 ,000 lb,σyield= 30 ,000 psi, and
E= 30 × 106 psi, formulate the problem as a parametric optimization problem for min-
imum volume design subject to buckling and yielding constraints. Assume the bars to be
pin connected for the purpose of buckling analysis. Indicate the procedure that can be
used for a graphical solution of the problem.


7.39 Minimizef (X)=(x 1 − 1 )^2 +(x 2 − 2 )^2
subject to
x 1 + 2 x 2 − 2 = 0


using the augmented Lagrange multiplier method with a fixed value ofrp=1. Use a
maximum of three iterations.
Free download pdf