An Extension of Lecture 01, Class Exercise 3: Production

Schedule Problem

Part 1: A Production Schedule Problem

Class Exercise 3: A Production Schedule Problem. A manufacturer of printed circuits has a stock
capacity of
• 200 resistors, 120 transistors and 150 capacitors
It required to produce two types of circuits:

• Type C1 requires 20 resistors, 10 transistors and 10 capacitors

• Type C2 requires 10 resistors, 20 transistors and 30 capacitors

The profit on type C1 circuits is £5 and that on type C2 circuits is £12. How many of each circuit
should be produced in order to maximise the profit?

You will not actually solve this problem yet, but show how it can be formulated using a
mathematical model. There are three vital stages in the formulation, namely,

(a) What are the unknowns?

(b) What are the constraints?
(c) What is the profit/cost to be maximised/minimised?

Sample Solution:

A tabular form as shown in Table 1 is usually useful for setting up a mathematical model for a
production schedule problem.

Table 1
Elements Types of circuits Stock capacity
C1 C2
Resistors 20 10 200
Transistors 10 20 120
Capacitors 10 30 150
Benefit £5 from producing each C1 £12 from producing each C2

(a) What are the unknowns?

Clearly, we want to determine the number of type C1 and type C2 circuits to be produced. So, we
define
x = number of type C1 circuits produced
y = number of type C2 circuits produced
(b) What are the constraints?
There are constraints associated with the total number of resistors, transistors and capacitors
available.
Registers: Since each type C1 requires 20 resistors, each type C2 requires 10 resistors, and there is
a total of 200 resistors available, then

20x + 10y ≤ 200,

Transistors: Similarly,

10x + 20y ≤ 120,

Capacitors: Similarly,

10x + 30y ≤ 150,

Nonnegative constraints: We must state the obvious (but nevertheless important) inequalities:
x ≥ 0, y ≥ 0.

(c) What is the profit?

Since each type C1 gives £5 profit and each type C2 gives £12 profit, the total profit is then

P = 5x +12y

Finally, we can summarise the problem as:

maximise P = 5x +12y
subject to 20x + 10y ≤ 200
10x + 20y ≤ 120
10x + 30y ≤ 150
x≥0
y ≥ 0.

Part 2: An Extension

(d) Draw a graph to show the region determined by all the constraints. Such a region is called feasible
region of the optimization problem.

(e) Prove, by applying the definition of convex set, that the region determined by all the constraints in the
optimization problem is a convex set.
Sample Solution
(d)
The feasible region is shown in figure 1.

20

5 (6,3) (4/3, 28/3)

3
Feasible Region x
6 10 12 15

Figure 1 Feasible region determined by the constraints for production schedule problem.

(e)
Let
 20 10   200 
x 
A  10 20  , b  120  , and z   
 
 y
10 30  150 

Using the above result, we can know that the area (region) formed by the following
constraints is a convex set:
Az ≤ b (1.1)

This is actually the example discussed in Lecture Note02, slide 9 (Fundamentals).

Note that the region defined by (1.1) is open (not closed or bounded). For convenience of
description, the symbol x used in the question was replaced by z. This is allowable.
Sample Solution #2 for (e)

An alternative approach to prove the area (region) formed by the following constraints is a convex
set:

 Az  b
 (1.2)
z  0

Let P1 = [x1, y1]T and P2 = [x2, y2]T be two arbitrary points in S, and  [0,1] . From the definition
of S, we have

20 x1  10 y1  200 2 x1  y1  20
10 x  20 y  120  x  2 y  12
 1  1
 
1 1
 (P1 is in S) (1.3)
10 x1  30 y1  150  x1  3 y1  15
 x1  0, y1  0  x1  0, y1  0

20 x2  10 y2  200 2 x2  y2  20
10 x  20 y  120  x  2 y  12
 2  2
 
2 2
 (P2 is in S) (1.4)
10 x2  30 y2  150  x2  3 y2  15
 x2  0, y2  0  x2  0, y2  0

Now, consider the new point P3 = λP1 + (1– λ)P2 =[ x3, y3]T. Writing it out in vector format, we
have,

x   x   x  (1   ) x2   x3 
P3   P1  (1   ) P2    1   (1   )  2    1   (1.5)
 y1   y2   y1  (1   ) y2   y3 

In order to prove that P3 is in S, we need to show

 2 x3  y3  20
 x  2 y  12
 3 3
 (1.6)
 3x  3 y3  15
 x3  0, y3  0

First of all, it is easy to show that x3 = λx1+(1- λ)x2 ≥ 0, and similarly y3 ≥ 0.

From (1.3) and (1.4), we have

2 x3  y3  2[ x1  (1   ) x2 ]  [ y1  (1   ) y2 ]
  (2 x1  y1 )  (1   )(2 x2  y2 )
(1.7)
   20  (1   )  20
=20
That is,
 2 x3  y3  20 (1.8)
Similarly,

x3  2 y3  12 (1.9)

x3  3 y3  15 (1.10)

Therefore, for any two points P1 = [x1, y1]T and P2 = [x2, y2]T , and  [0,1] , the new point z = P3 =
λP1 + (1– λ)P2 =[ x3, y3]T possesses the following properties

 Az  b
 (1.11)
z  0
In other words, the combination z = λP1 + (1– λ)P2 =[ x3, y3]T is still in S. Therefore, from the
definition of convex sets, the set S defined by the following constraints is convex:

20x + 10y ≤ 200

10x + 20y ≤ 120
10x + 30y ≤ 150
x ≥ 0, y ≥ 0.