Scribd
Upload
68 views

HW 7 Solutions

The document provides details for two homework problems (A and B) involving optimization of electric power systems. Problem A involves setting up and solving an optimal power flow problem as a linear program, with the objective of minimizing costs. It provides the system diagram, constraints, and asks the student to provide the optimal solution variables and objective value for the original problem and a variation where one line flow is constrained. Problem B involves using branch-and-bound to solve two integer programming problems - problem B1 involves maximizing a linear objective subject to two constraints, and problem B2 involves maximizing another linear objective with two constraints and an integrality constraint on one variable. The student is asked to solve each problem by successive LP

Uploaded by

karen dejo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
68 views

HW 7 Solutions

The document provides details for two homework problems (A and B) involving optimization of electric power systems. Problem A involves setting up and solving an optimal power flow problem as a linear program, with the objective of minimizing costs. It provides the system diagram, constraints, and asks the student to provide the optimal solution variables and objective value for the original problem and a variation where one line flow is constrained. Problem B involves using branch-and-bound to solve two integer programming problems - problem B1 involves maximizing a linear objective subject to two constraints, and problem B2 involves maximizing another linear objective with two constraints and an integrality constraint on one variable. The student is asked to solve each problem by successive LP

Uploaded by

karen dejo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

EE/Econ 458: HW 7, Due Wednesday Oct 26

A. You solved the following optimization problem in HW6. Using


the same system you analyzed in problem 1, shown below, set up
the optimal power flow as a linear program. Assume the objective
function is the same as used in the example in class when we
investigated the case of demand bidding, i.e.,
Z  1307 Pg1  1211Pg 2  1254Pg 4  1300Pd 2  1200Pd 3 . Also, assume (as in
HW6) each generator has a lower limit of 100 MW and an upper
limit of 300 MW, which will be (in per unit):
1  Pg1  3
1  Pg 2  3
1  Pg 3  3
and the loads are constrained as follows:
1  Pd 2  2
2  Pd 3  3

Pg2=2pu
Pg1

1 2

y12 =-j10
y14 =-j10 y13 =-j10
Pd2=1pu y23 =-j10
y24 =-j8

y34 =-j10
4 3

Pg4=1pu Pd3=4pu

Problem A1: Provide the objective function and all primal decision
and auxiliary variables and dual variables at the optimal solution.
Also identify the settlement, i.e., identify how much each load pays
and how much each generator pays.
Solution:
Objective: -17.0$
Primal decision and auxiliary variables:
pg1 1.000000
pg2 2.000000

1
pg4 1.000000
pd2 2.000000
pd3 2.000000
pb1 -0.013889
theta4 0.001389
pb2 0.263889
theta2 -0.026389
pb3 0.486111
theta3 -0.075000
pb4 0.763889
pb5 0.750000
pb6 -0.222222
Dual variables:
c8 1211.000000
c9 1211.000000
c10 1211.000000
c11 1211.000000
c12 -96.000000
c16 -43.000000
c19 -89.000000
c20 -11.000000
All other dual prices in the range 1-33 are 0.
Settlement:

Problem A2: Now constrain the flow on branch 3 to Pb3=P23=0.45


pu and resolve. Provide the objective function and all solution
variables and dual variables at the optimal solution. Also identify
the settlement, i.e., identify how much each load pays and how
much each generator pays. Finally, compute the congestion charges
as the difference between total load payments made and total
generator payments received, and using the dual variable of the
constrained branch.
Objective: -11.41$
Primal decision and auxiliary variables:
pg1 1.000000

2
pg2 1.870000
pg4 1.130000
pd2 2.000000
pd3 2.000000
pb1 -0.050000
theta4 0.005000
pb2 0.300000
theta2 -0.030000
pb3 0.450000
theta3 -0.075000
pb4 0.800000
pb5 0.750000
pb6 -0.280000
All other variables in the range 1-15 are 0.
Dual variables:
c1 0.000000
c4 -154.800000
c8 1251.850000
c9 1211.000000
c10 1290.550000
c11 1254.000000
c12 -55.150000
c19 -89.000000
c20 -90.550000
c27 -154.800000
All other dual prices in the range 1-33 are 0.
Settlement:

B. Use branch-and-bound to solve problems B1 and problem B2.


For both problems:
a. Solve them using successive LP-relaxations, where each LP-
relaxation is solved using the CPLEX (or Matlab) LP-solver.
b. Solve them using the CPLEX MIP-solver.

Problem B1:

3
Max z = 5x1 + 2x2
s.t. 3x1 + x2 ≤ 12
x1 + x2 ≤ 5
x1, x2, ≥ 0; x1, x2 integer

Solution to B1:

a. Using successive LP-relaxations:

First solve as a relaxed LP:


P1: Max z=5x1+2x2
s.t. 3x1+x2≤ 12
x1 + x2 ≤ 5
 z*=20.5, x1*=3.5, x2*=1.5

P2: Max z=5x1+2x2


s.t. 3x1+x2≤ 12
x1 + x2 ≤ 5
x1>=4
x2 ≥ 0
z*=20, x1*=4, x2*=0
This solution is integer, so z*=20 becomes our best feasible
solution.

P3: Max z=5x1+2x2


s.t. 3x1+x2≤ 12
x1 + x2 ≤ 5
x1<=3
x2 ≥ 0
z*=19, x1*=3, x2*=2
The objective function value is not as good as that of our best
feasible solution and therefore we may stop pursuing this branch.
Let’s go back and force x2 to be integer.

4
P4: Max z = 5x1 + 2x2
s.t. 3x1 + x2 ≤ 12
x1 + x2 ≤ 5
x2 ≥ 2
x1 ≥ 0
x1, x2, ≥ 0
z*=19, x1*=3, x2*=2
The objective function value is not as good as that of our best
feasible solution and therefore we may stop pursuing this branch.

P5: Max z = 5x1 + 2x2


s.t. 3x1 + x2 ≤ 12
x1 + x2 ≤ 5
x2 ≤ 1
x1 ≥ 0
x1, x2, ≥ 0
z*=20.3, x1*=3.667, x2*=1
The objective function value is better than that of our best feasible
solution, but this solution is not integer. Therefore we need to
continue pursuing this branch. Let’s force x1=3.

P6: Max z = 5x1 + 2x2


s.t. 3x1 + x2 ≤ 12
x1 + x2 ≤ 5
x2 ≤ 1
x1 ≤ 3
x1, x2, ≥ 0;
z*=17, x1*=3, x2*=1
The objective function value is not as good as that of our best
feasible solution and therefore we may stop pursuing this branch.
Let’s force x1=3.:

5
P7: Max z = 5x1 + 2x2
s.t. 3x1 + x2 ≤ 12
x1 + x2 ≤ 5
x2 ≤ 1
x1 ≥ 4
x1, x2, ≥ 0;
z*=20, x1*=4, x2*=0
Which is the same solution as that obtained in P2. All possible
branches have been considered, therefore the solution to P2 becomes
the solution to the problem: z*=20, x1*=4, x2*=0.

b. Using CPLEX MIP solver:

z*=20, x1*=4, x2*=0.

Problem B2:
Max z = 3x1 + x2
s.t. 5x1 + 2x2 ≤ 10
4x1 + x2 ≤ 7
x1, x2,≥ 0; x2 integer

Solution to Problem B2:


Max z = 3x1 + x2
s.t. 5x1 + 2x2 ≤ 10
4x1 + x2 ≤ 7
x1, x2,≥ 0; x2 integer

a. Using successive LP-relaxations:

First solve as a relaxed LP:


P1: Max z = 3x1 + x2

6
s.t. 5x1 + 2x2 ≤ 10
4x1 + x2 ≤ 7
x1, x2,≥ 0;
z*=5.667, x1*=1.333, x2*=1.667

Force x1=1.

P2: Max z = 3x1 + x2


s.t. 5x1 + 2x2 ≤ 10
4x1 + x2 ≤ 7
x1≤1
x1, x2,≥ 0;
z*=5.5, x1*=1, x2*=2.5

Force x2=2.

P3: Max z = 3x1 + x2


s.t. 5x1 + 2x2 ≤ 10
4x1 + x2 ≤ 7
x1≤1
x2≤2
x1, x2,≥ 0;
z*=5.0, x1*=1, x2*=2

This is a feasible solution (all integer) and therefore z*=5 represents


our best bound found so far. Now let’s branch the other way from
P2 by forcing x2=3.

P4: Max z = 3x1 + x2


s.t. 5x1 + 2x2 ≤ 10
4x1 + x2 ≤ 7

7
x1≤1
x2≥3
x1, x2,≥ 0;
z*=5.4, x1*=0.8, x2*=3

P4 is not feasible but it has a better bound than our best bound so far
(of 5.0). So we need to branch further on P4. Force x1 to 1.

P5: Max z = 3x1 + x2


s.t. 5x1 + 2x2 ≤ 10
4x1 + x2 ≤ 7
x1≥1
x2≥3
x1, x2,≥ 0;
This one is infeasible. Go the other way by forcing x1 to 0.

P6: Max z = 3x1 + x2


s.t. 5x1 + 2x2 ≤ 10
4x1 + x2 ≤ 7
x1≤0
x2≥3
x1, x2,≥ 0;
z*=5.0, x1*=0, x2*=5
This is a feasible solution and it has objective function value equal
to our best bound so far (P3).

So now go back to P1 and branch the other way by forcing x1=2.

P7: Max z = 3x1 + x2


s.t. 5x1 + 2x2 ≤ 10
4x1 + x2 ≤ 7
x1≥2

8
x1, x2,≥ 0;
This problem is infeasible, and so we are done. There are two
answers:
From P3: z*=5.0, x1*=1, x2*=2
From P6: z*=5.0, x1*=0, x2*=5

b. Using MIP solver:

z*=5, x1*=0, x2*=5.

You might also like