Operations Research Dynamic programming

Dynamic programming
Introduction:
Dynamic programming is a mathematical technique dealing with the optimization of
multistage decision process. The word ‘dynamic’ means to the situations of change in
several stages, such as every week, every day, or every month etc. and programming is
the term used in the mathematical sense of selecting an optimum allocation of resource.
The dynamic programming determines the optimum solution to an n-variable problem by
decomposing it into n stages with each stage constituting a single variable sub problem.
Recursive nature of dynamic programming:
Computations in DP are done recursively in the sense that the optimum solution of one
sub problem is used as an input to the next sub problem. By the time the last sub problem
is solved, the optimum solution for the entire problem is at hand.
Principle of optimality:
The concept of D.P is largely based upon the principle of optimality due to “Bell man”,
viz.
“An optimal policy has the property that what ever the initial state and initial
decision are, the remaining decisions must constitute an optimal policy with regard to the
state resulting from the first decision.
The principle of optimality implies that given the initial state of a system, an
optimal policy for the subsequent stages does not depend upon the policy adopted at the
preceding stages. That is the effect of a current decision on any of the policy decisions of
the preceding stages need not be taken into account at all. It is usually referred to as the
Monrovians Property of Dynamic Programming.
Salient features and their definitions of DPP
1. Recursive function: a DPP is solved by an objective function which is recurring in
nature. This recursive approach can be applied in two directions.
a) Forward recursive approach: In this approach the main problem is attacked by
dividing it from initial stage to final stage.
b) Backward recursive approach: In this approach the problem is dealt by
subdividing the problem from last to first and then solving from final to initial
stages.

1
Operations Research Dynamic programming

2. Stage: A large problem is decomposed into smaller and sequential sub problems is
referred to as a ‘stage’, where a decision is called for. Thus each stage can be
considered to have a starting and ending. The ending of one stage will be the
beginning of its next stage.
Suppose an engineering student has to acquire good marks in his degree after
studying for four years. Each year is considered to be a stage. When a student
completes his first year successfully, he is said to have ended first stage (and started
his second stage) having 1/4th of his large problem is completed. Thus he moves to
optimize his returns in second stage i.e. 2nd year with similar objective function, but
this includes the results of first year and completion of his second stage he acquires
2/4th of problem solved. Similarly he goes to 3 rd stage, finishing 3/4th that should
include the results of first, 2nd year. Finally, after completing his fourth year, he
completes his total problem that includes results of his previous three stages.

3. State: any stage just before starting or just after completion will have certain status
called the “state” and the variable that links up two stages is called state variable.
4. Return function: The decision of a stage described in the form of an algebraic
equation with state variables to explain the worth of benefit (or cost) is known as
return function or transformation equation.
5. Decision or policy variable: a decision is made at each stage. The decision or policy
variable transforms the state variable from any given stage into a state associated with
the next stage. For example, at any stage in the salesmen allocation problem, a policy
would be the allocation of some specific number of the available salesmen to the
territory represented by that stage.

2
Operations Research Dynamic programming

6. Optimal policy: a policy which optimizes the value of a criterion, return or objective
function is called an optimal policy starting in any given state of any stage; the
optimal policy depends only upon that state and not upon how it was reached. In
accordance with the principle of optimality, the optimal return function is found by
optimizing the immediate return (cost) from stage n plus the known optimal return
(cost) from the previous stage n-1.
The general relationship between state, stage and policy (or decision) can be
illustrated as shown in figure.

On the basis of the principle of optimality the multi-stage problem is

solved by first solving the problem in the last stage, then working backward and
solving the problem in the next-to-last stage and so on until first stage problem is
solved. This procedure is called backward recursion.
7. State transformation: Sn-1= Tn (Sn, Dn) is a function that tells what the state of the
process will be at the start of the next state (stage s n-1) as a function of Sn (the state at
the start of stage n ) and D n (the decision at stage n). Here, T n is the state
transformation function and its form depends upon the particular problem to be
solved.
Applications of D.P
1. shortest route problem
2. cargo loading
3. work force size model
4. equipment replacement
5. inventory models
6. investment models
7. application in production

3
Operations Research Dynamic programming

8. application in linear programming

9. application in reliability
10. application in continuous systems
11. design of continuous beams
12. optimal layout of truss
13. optimal design of gear train
14. design of a minimum cost drainage system

Characteristics of dynamic programming:

1. The problem can be sub divided into stages with a policy decision at each stage. A
stage is a device to sequence the decisions. That is, it decomposes a problem into
sub-problems, Such that an optimal solution to the problem can be obtained from
the optimal solutions to the sub problems.
2. Every stage consists of a number of states associated with it. The states are the
different possible conditions in which the system may find itself at that stage of
the problem.
3. Decision at each stage converts the current stage into state associated with the
next stage.
4. The state of the system at a stage is described by a set of variables, called state
variables.
5. When the current state is known, an optimal policy for the remaining stages is
independent of the policy of the previous ones.
6. To identify the optimum policy for each state of the system, a recursive equation
is formulated with n stages remaining, given the optimal policy for each state with
(n-1) stages left.
7. using recursive equation approach each time the solution procedure moves
backward stage by stage for obtaining the optimum policy of each state for that
particular stage, till it attains the optimum policy beginning at the initial stage.

4
Operations Research Dynamic programming

Dynamic programming algorithm:

1. Identity the decision variables and specify objective function to be optimized under
certain limitations, if any.
2. Decompose the given problem into a number of smaller sub-problems (or stages).
Identify the state variables at each stage and write down the transformation function
as a function of the state variables and decision variables at the next stage.
3. Write down a general recursive relationship for computing the optimal policy. Decide
whether forward or backward method is to follow to solve the problem.
4. Construct the appropriate stages to show the required values of the return function at
each stage.
5. Determine the overall optimal policy or decisions and its value at each stage. There
may be more than one such optimal policy.

Linear programming as a case of Dynamic programming:

A linear programming problem with n decision variables and m constraints can be
considered as an n-stage dynamic programming problem with m stage variables.

Maximize f(X1, X2, X3------Xn) =

Subject to: i=1, 2, -----------m

Xj

This is an n-stage decision problem where the value of the decision variable must be
determined at stage j.
The right hand side of the constraints bi can be treated as m types of resources to
be allocated among different kinds of activities Xj (different kinds of activities may be
number of castings produced, number of forgings produced, number of finished
components produced), bi may be available machines, available time. The constant Cj
represents the profit per unit bi needed for 1 unit of jth activity Xj (E.g.: the amount of
material required to produce one casting).
By denoting the optimal value of the composite objective function over n-stages as fn*.

We can state the problem as

5
Operations Research Dynamic programming

Find:

fn*= fn*(b1,b2,b3, --------bm)=

Such that

Xj

The recurrence relationship applied to this problem yields.

fi*( )=
Where are the sources available for allocation at stage i
a1iXi, a2iXi, ---------amiXi are the resources allocated to the activity Xi.
Resources available for allocation to the activity i-1,
indicates the maximum value that Xi can take without violating any f the constraints.

Since any value larger than would violate at least one constraint. Thus at the i th stage,
the optimal values Xi* and fi* can be determined as functions of
Finally at the nth stage, since the values of are known to be b1, b2, b3,
--------bm, respectively, we can determine X n* and fn*. Once Xn* is known, the remaining
values Xn-1*, Xn-2*, ------X1* can be determined by retracing the sub optimization steps.

Problems on LPP by using dynamic programming

1. Maximize f(X1, X2) = 50 X1+100X2

Subject to: 10 X1+5 X2
4 X1+10X2

6
Operations Research Dynamic programming

X1+1.5X2
X1, X2
Solution:
Since n=2; m=3; this problem can be considered as a two stage problem with three state
parameters.
Stage 1: max f1 (
X1 is a non-negative value that satisfies the side constraints.
10X1
Here and hence the maximum value
that X1 can assume is given by
=X1*=min

=50 min
Stage II: maximum value of f2

Max f2 (

are the resources available for allocation at stage II, which are equal to 2500,
2000, 450 respectively.
The maximum value of X2 can assume without violating any constraint is given by

=
Thus the recurrence relation is restated as

Maxf2(

Since min

=max
=max (2500-25X2) =21,875 at X2= 125

7
Operations Research Dynamic programming

Hence
f2*(2500, 2000, 450) =21,875 at X2*= 125.
X1*= min
= min (187.5, 187.5, 262.5) =187.5
X1*=187.5; X2*= 125; f max*= 21,875.

2. Maximize Z= 10 X1+ 30X2

Subject to: 3X1+6 X2
12X2
X1, X2
Solution:
Since n=2; m=2; this problem can be considered as a two stage problem with 2 state
parameters.
Stage II: maximum value of f2:

Max f2 (

Where X2 is a non-negative value that satisfies the side constraints.

6X2
Here and hence the maximum value that X2 can assume is
given by
=X2*=min

f2*(168-3X1, 20) =30X2*= 30 min

Stage 1:

Maximum value of X1:

Max f1 (

are the resources available for allocation at stage I, which are equal to 56, 20.

8
Operations Research Dynamic programming

Max f1 (

Since min

X1*=16; X2*= min =min [20, 20] =20

Z max= 760

X1*=16; X2*=20

3. Max Z= 2 X1+ 5X2

Subject to: 2X1+ X2
2X2
X1, X2
Solution:
Since n=2; m=2; this problem can be considered as a two stage problem with 2 state
parameters.

Stage II: maximum value of f2:

Max f2 (

Where, X2 is a non-negative value that satisfies the side constraints.

X2
Here and hence the maximum value that X2 can assume is given
by
=X2*=min

9
Operations Research Dynamic programming

f2* =5X2*= 5 min

Stage 1:

Maximum value of X1:

Max f1 (

are the resources available for allocation at stage I, which are equal to 21.5, 23.

Max f1 (

Since min

X1*=10; X2*= min =min [23, 23] =23

Zmax= 135

X1*=10; X2*=23
4. Max Z= X1+ 9X2
Subject to: 2X1+ X2
X2
X1, X2
Solution:
Since n=2; m=2; this problem can be considered as a two stage problem with 2 state
parameters.
Stage II: maximum value of f2:

Max f2(

Where X2 is a non-negative value that satisfies the side constraints.

10
Operations Research Dynamic programming

X2
Here and hence the maximum value that X2 can assume is given
by
=X2*=min

f2* =9X2*= 9 min

Stage 1:

Maximum value of X1:

Max f1 (

are the resources available for allocation at stage I, which are equal to 12.5, 11.

Max f1 (

Since min

X1*=7; X2*= min =min [11, 11] = 11.

Z max= 106

X1*=7; X2*=11

Cargo loading:
1) Find number of each of three items to be included in a package so that value of
package will be maximum. Total weight of package must not exceed 5 kg’s.

11
Operations Research Dynamic programming

Item Weight in kg’s Value in Rs

1 1 30
2 3 80
3 2 65

Solution:
The given problem can be formulated as follows:
Let X1, X2 and X3 be the number of items of each type.
Maximize f(X) =30X1+80X2+65X3 (The total value is to be maximized)
Subject to: X1+X2+X3
X1, X2, and X3
We have to determine many units of three types of items are to be loaded. So it is a three
stage dynamic programming problem.
Let Xj (j=1,2,3) denote three decisions. Let fj (Xj) denote the value of the optimal
allocation for the three types of items.
If fj (S, Xj) be the value associated with the optimum solution f j (S), (j=1, 2, 3, ------n)
then we have

f1*(S) = and fj*(S) = where j=1, 2, 3

Where Pj (Xj) denotes the expected value obtained from allocation of X j units of weight to
the item j.
Now for stage I problem:

f1*=

Where the largest value of X1=

Hence we have six alternatives i.e., 0, 1, 2, 3, 4 and 5 for X 1 we have the following
computations.
Stage I:

X1 Value of 30X1 Optimum solution

S 0 1 2 3 4 5 f1*(S) X1*
0 0 - - - - - 0 0

12
Operations Research Dynamic programming

1 0 30 - - - - 30 1
2 0 30 60 - - - 60 2
3 0 30 60 90 - - 90 3
4 0 30 60 90 120 - 120 4
5 0 30 60 90 120 150 150 5

Stage II: The largest value of X2 is 5/3=1 and thus we have two alternatives i.e. 0 and 1
for X2.

f2*(S) =

X2 Value of Optimum solution

S 0 1 f2*(S) X2*
0 0+0=0 - 0 0
1 0+30=30 - 30 0
2 0+60=60 - 60 0
3 0+90=90 80+0=80 90 0
4 0+120=120 80+30=110 120 0
5 0+150=150 80+60=140 150 0

Stage III:
The largest value of X3 is 5/2= 2 and we have 0, 1 and 2 as the alternatives for X3
And

f3*(S) =

X3 Value of Optimum
S solution
0 1 2 f3*(S) X3*
0 0+0=0 - - 0 0
1 0+30=30 - - 30 0
2 0+60=60 65+0=65 65 1
3 0+90=90 65+30=95 95 1
4 0+120=120 65+60=125 130+0=130 130 2
5 0+150=150 65+90=155 130+30=160 160 2

For the given total weight W=5 the optimum solution is, X1*=1; X2*= 0; X3*=2;
Maximum value of Rs.160

13
Operations Research Dynamic programming

2. A 4-ton vessel loaded with one or more of three items. The following table gives
the unit weight, wi, in tons and the unit revenue in thousands of dollars, r i for item i.
how should the vessel be loaded to maximize the total return?

Item i Wi Ri
1 2 31
2 3 47
3 1 14

Solution:
Stage III

f3*=

Where the largest value of X3=4/1=4

Hence we have five alternatives i.e. 0, 1,2,3,4 for X3
We have the following computations.
Stage III:
X3 Value of 14X3 Optimum solution
S 0 1 2 3 4 f3*(S) X3*
0 0 - - - - 0 0
1 0 14 - - - 14 1
2 0 14 28 - - 28 2
3 0 14 28 42 - 42 3
4 0 14 28 42 56 56 4

Stage II:
The largest value of X2=4/3=1 and thus we have two alternatives i.e. 0 and 1 for X2

f2*(S) =

X2 Value of Optimum solution

S 0 1 f2*(S) X2*
0 0+0=0 - 0 0
1 0+14=14 - 14 0
2 0+28=28 - 28 0
3 0+42=42 47+0=46 47 1
4 0+56=56 47+14=61 61 1

14
Operations Research Dynamic programming

Stage I:
The largest value of X1=4/2=2 and we have 0, 1 and 2 as alternatives for X1 and
f1*(S) =

X1 Value of Optimum
S solution
0 1 2 f1*(S) X1*
0 0+0=0 - - 0 0
1 0+14=14 - - 14 0
2 0+28=28 31+0=31 31 1
3 0+47=47 31+14=45 47 0
4 0+61=61 31+28=59 62+0=62 62 2

X1*=2; X2*=0; X3*=0; Z max=62;

Shortest route problem:

1. Select the shortest highway route between two cities. The network in fig: provides
the possible routes between the starting city at node 1 and destination city at node 7.

Solution:
To solve the problem by D.P, we first decompose it into stages as delineated by the
vertical dashed line in figure. Next, we carry out the computations for each stage
separately.

15
Operations Research Dynamic programming

The general idea is to compute the shortest (cumulative) distances to all the terminal
nodes of a stage and then use these distances as input data to the immediately succeeding
stage. Stage 1 includes three nodes (2, 3, 4) and its computations are simple.
Stage I:
Shortest distance to node 2= 7 miles (from node 1)
Shortest distance to node 3= 8 miles (from node 1)
Shortest distance to node 4= 5 miles (from node 1)
Stage II:
Stage 2 has two end nodes (5 and 6) considering node 5 first, we see from fig. that there
are three possible routes to reach node 5-namely [(2, 5), (3, 5) and (4, 5)]. This
information together with the shortest distances to nodes 2, 3 and 4, determines the
shortest (cumulative) distance to node 5 as
Shortest distance to node 5

= [shortest distance to node i + distance from node i to node 5].

=min =12 (from node 4)

Similarly for node 6 we have:

Shortest distance to node 6 = (from node 3)

Stage III:
Shortest distance to node 7:=min
Therefore, shortest distance from node 1 to node 7 = 21 miles.

16
Operations Research Dynamic programming

From stage III, the node 7 is linked to node 5. Next node 5 is linked to node 4. Finally
node 4 is linked to node 1. Thus the shortest route is defined as

Allocation:
1. A member of a certain political party is making plans for his election to the
parliament. He has received the service of six volunteer workers and wishes to
assign them to three districts in such way as to maximize their effectiveness. He
feels that it would be inefficient to assign no worker to more than one district,
but he is willing to assign no worker to any one of the district if they can
accomplish in other districts. The following table gives the estimated increase in
the number of votes in his favour in each district if it were allocated various
numbers of workers.
Number of Districts
workers 1 2 3
0 0 0 0
1 25 20 33
2 42 38 43
3 55 54 47
4 63 65 50
5 69 73 52
6 74 80 53

How many of the six workers should be assigned to each of the three districts in
order to maximize total estimated increase in the number of votes in his favour.
Solution:
Maximize Z= V1X1+V2X2+V3X3
Subject to: X1+ X2+ X3=6
Recursive relation:
f1(S, X1) = V1X1
If fj (S, Xj) be the profit associated with the optimum solution fj*(S), j=1, 2, 3 then

f1*(S, Xj) =

Stage-3
S f3*(S) X3*

17
Operations Research Dynamic programming

0 0 0
1 33 1
2 43 2
3 47 3
4 50 4
5 52 5
6 53 6

Stage-2:
X2 Value of f2(S,X2)=V2X2+f3*(S-X2) Optimum
solution
S
0 1 2 3 4 5 6 f2(S) X2*

0 0+0=0 - - - - - - 0 0
1 0+33=33 20+0=20 - - - - - 33 0
2 0+43=43 20+33=53 38+0=38 - - - - 53 1
3 0+47=47 20+43=63 38+33=71 54+0=54 - - - 71 2
4 0+50=50 20+47=67 38+43=81 54+33=87 65+0=65 - - 87 3
5 0+52=52 20+50=70 38+47=85 54+43=97 65+33=98 73+0=73 - 98 4
6 0+53=53 20+52=72 38+50=88 54+47=101 65+43=108 73+33=106 80+0=80 108 4

Stage 1:
X1 Value of f2(S,X2)=V2X2+f3*(S-X2) Optimum
S solution
0 1 2 3 4 5 6 f1(S) X1*
0 0 - - - - - - 0 0
1 33 25+0=20 - - - - - 33 0
2 53 25+33=58 42 - - - - 58 1
3 71 25+53=78 42+33=75 55 - - - 78 1
4 87 25+71=96 42+53=95 55+33=88 63 - - 96 1
5 98 25+87=112 42+71=113 55+53=108 63+33=96 69 - 113 2
6 108 25+98=123 42+87=126 55+71=116 63+53=116 69+33=102 74 129 2
X1*= 2; X2*=3; X3*=1;

18
Operations Research Dynamic programming

2. The owner of a chain of four grocery stores has purchased six crates of fresh
strawberries. The estimated probability distribution of potential sales of the
strawberries before spoilage differs among the four stores. The following table
gives the estimated total expected profit at each store, when it is allocated
various numbers of crates.
For administrative reasons, the owner does not wish to split crates between
stores. However, he is willing to distribute Zero crates to any of his stores.
Find the allocation of six crates to four stores as to maximize the expected profit.
Store
1 2 3 4
Number of crates

0 0 0 0 0
1 4 2 6 2
2 6 4 8 3
3 7 6 8 4
4 7 8 8 4
5 7 9 8 4
6 7 10 8 4

Solution:
Let Xj be the number of crates allocated at the jth stage j=1, 2, 3, 4;
Pj (Xj) = expected profit from allocation of Xj crates to store j.
Now the problem can be formulated as LPP as follows
Max Z = P1 (X1) + P2 (X2)+ P3 (X3)+ P4 (X4)
Subject to: X1+ X2+ X3+ X4=6 and
X1, X2, X3, X4
Recursive equations:
f1(S, X1) = P1 (X1)
if fj* (S, Xj) denotes the profit associated with the optimum solution f j* (S) (j=1,2,3,4)

then f1* (S) =

Therefore the recurrence relation is obtained as

fj (S, Xj) = Pj (Xj) + fi+1 (S-Xj) where j=1, 2, 3, 4;

fj* (S) =

19
Operations Research Dynamic programming

Stage 4:
S f4* X4*
0 0 0
1 2 1
2 3 2
3 4 3
4 4 3, 4
5 4 3, 4, 5
6 4 3, 4, 5, 6

X4 Value of P4 (X4) Optimum solution

S 0 1 2 3 4 5 6 f4(S) X4*
0 0 - - - - - - 0 0
1 0 2 - - - - - 2 1
2 0 2 3 - - - - 3 2
3 0 2 3 4 - - - 4 3
4 0 2 3 4 4 - - 4 3, 4
5 0 2 3 4 4 4 - 4 3, 4, 5
6 0 2 3 4 4 4 4 4 3, 4, 5, 6

Stage 3:

X3 Value of P3(X3)+ f4*(S-X3) Optimum solution

S 0 1 2 3 4 5 6 f3(S) X3*
0 0 - - - - - - 0 0
1 2 6 - - - - - 6 1
2 3 8 8 - - - - 8 1, 2
3 4 9 10 8 - - - 10 2
4 4 10 11 10 8 - - 11 2
5 4 10 12 11 10 8 - 12 2
6 4 10 12 12 11 10 8 12 2, 3

Stage 2:

X2 Value of P2(X2)+ f3*(S-X2) Optimum solution

S 0 1 2 3 4 5 6 f2(S) X2*

0 0 - - - - - - 0 0
1 6 2 - - - - - 6 1
2 8 8 4 - - - - 8 0, 1
3 10 10 10 6 - - - 10 0, 1, 2
4 11 12 12 12 8 - - 12 1, 2, 3
5 12 13 14 14 14 9 - 14 2, 3, 4
6 12 14 15 16 16 15 10 16 3, 4

20
Operations Research Dynamic programming

Stage 1:

X1 Value of P1(X1)+ f2*(S-X1) Optimum solution

S 0 1 2 3 4 5 6 f1(S) X1*

6 16 18 18 17 17 13 7 18 1, 2

From above computations it is observed that the maximum profit of Rs.18 can be
obtained by choosing the following eight alternative solutions that each row must be 6.
Store 1 :X1 Store 2 : X2 Store 3: X3 Store 4: X4
1 2 2 1
1 3 1 1
1 3 2 0
1 4 1 0
2 1 2 1
2 2 1 1
2 2 2 0
2 3 1 0

1. Minimize Z =

Subject to: and

Solution:
Since the decision variables are so the given problem is a three stage
problem.

21
Operations Research Dynamic programming

The functional (recurrence) relation is

Since the function attains its minimum value at y2=

Since
Since minimum value of the function occurs
y3=5; y1+y2= 10; y1= 5;
Hence the optimal policy (5, 5, 5) with f3 (15) =75.

2. Maximize Z=

Subject to: and are positive integers.

Solution:
Since the decision variables are . So the given problem is a three stage
problem.

Stage 1:
S1= y1 where y1 can vary from 1 to 4

22
Operations Research Dynamic programming

with y1 integer

y1 1 2 3 4
f1(S1) 1 4 9 16

Stage 2:
With y1, y2 integers
y1 1 2 3 4
4 2 - 1
y2=

y1 1 2 4
f1(S1) 1 4 16
y2 y22
1 1 2 5 17
2 4 5 8
4 16 17

The optimal values of f2(X2) are

S2 1 2 4
f2(S2) 5 4 17

Stage 3:

=Max

S2 1 2 4
f2(S2) 2 5 17

23
Operations Research Dynamic programming

y3 y32
1 1 3 6 18
2 4 6 9
4 16 18

The optimal value of f3 (S3) are

S3 1 2 4
f3(S3) 3 6 18

For S3= 4; optimal Z = 18;

y1=1; y2= 4; y3=1;
3. Max; y1y2y3
Subject to: y1+y2+y3 =5;
Solution: this dynamic programming problem can be treated as a three stage problem
with state variables Si and return fi (si) such that
At stage 3:

At stage 2:

At stage1:

And the returns

Since y1 = S2-y2

Differentiating w.r.t y2 and equating to Zero.

24
Operations Research Dynamic programming

f2 (S2) =

f3 (S3) =

Differentiating w.r.t y3 and equating to Zero.

y3= and y1 , y2 and maximum value y1*y2* y3 =

Assignment problems

1. Max Z= 3 X1+ 4X2

Subject to: X1
X2
2X1+ X2
X1, X2
2. Max Z=4 X1+12X2
Subject to: 2 X1+6 X2
6 X1+2X2
X1, X2
3. Max Z=8 X1+7X2
Subject to: 2 X1+ X2
5 X1+2X2
X1, X2

25
Operations Research Dynamic programming

4. Max Z= 50 X1+100X2
Subject to: 2 X1+3 X2
X1+3X2
X1+X2
X1, X2 Sol: X1=6, X2=12; Zmax=60

5. Max Z=3 X1+7X2

Subject to: X1+4 X2
X2
X1, X2 Sol: X1=8, X2=0; Zmax=24

6. Max Z= 50x1+80x2
Subject to: x1 , x2 , 5x1+6x2 , x1+2x2 , x1,x2
Sol: Z= 7,000
x1=60, x2= 50.
7. Find out the minimum distance from city A to city B from the fig as shown in below:

Solution:
Total distance =3+3+6+4=16 km’s.
Route:

8. Truck can carry a total of 10 tons of commodity. Three types (A, B & C) of
commodities are to be carried. These commodities have the characteristics as shown in
the following table:

Commodity unit weight profit per

ton (Rs)
A 4 100
B 5 130
C 3 80

26
Operations Research Dynamic programming

9. A ship is to be loaded with certain items. Each unit of item ‘i’ has a weight wi and a
value vi (i=1, 2, 3). The maximum cargo weight permitted is W. using the following
tables; determine the most valuable cargo load which will not exceed the maximum
permissible weight.
i) When W=10;
i Wi vi
1 5 4
2 8 10
3 3 6

ii) When W= 1000

i Wi vi
1 495 220
2 500 750
3 510 1012

iii) When W=10 tons

commodity Unit weight (tons) Profit per ton (Rs.)
A 4 100
B 5 130
C 3 80

10. A company makes three types of product A, B and C. the raw material required and
the cash return for each product are shown below in the table. The total number of
boxes of raw materials available for the given period is 10. The company wants to
select a product mix that maximizes the total cash return. Since a fractional unit of a
product is meaningless, the product must be measured in integer values.
Product Boxes of raw material required for per unit of Profit per unit
product (Rs.)
A 4 10
B 3 7
C 2 4

27
Operations Research Dynamic programming

11. A student has to take examination in three courses. He has only 5 days left. Table
below indicates the relationships between the expected grade in a particular course
and the block of time spent studying it.
Block of time Course
available (days)
A B C
0 0 0 0
1 4 3 2
2 6 5 4
3 7 6 5
4 9 7 7
5 10 8 9

12. Mr. X wishes to invest Rs.50, 000 among three possible investment projects. He
could invest all, part or none of his Rs.50, 000 in any one of the project, but he has to
invest in units of Rs.10, 000. the expected return, which depends on the amount of
money invested is shown in the table given below:

Amount of Expected return from project (Rs.)

investment(Rs.)
A B C
0 0 0 0
10 8 9 12
20 15 14 16
30 21 25 24
4 26 30 28
5 30 34 32

University question papers:

1. a) What do you mean by “stage” and “state” in dynamic programming?
b) XYZ road ways has four types of packages A, B, C and D to be carried out in their
parcel van. The bulk density of each package is different, as per the company’s rules
the packages fall under different categories of freight classification and therefore, the
revenue per unit of each package also varies. Data regarding the weights and the
expected revenue for each package are as follows.
Type of package Weight per unit (Kg) Expected revenue (Kg)
A 2000 200
B 4000 425

28
Operations Research Dynamic programming

C 5000 550
D 3000 350
Determine the number of units of each package that would maximize the revenue,
given that the capacity of the van is limited to 10,000 Kgs. Use dynamic
programming approach.
2. a) What are the situations warranting use of dynamic programming.
b) Six units of capital is available to invest in four business ventures. The returns
from each unit of investment in all the four ventures are given in the table below.
Find how the capital should be allocated to business proposals in order to maximize
profit using dynamic programming.

Expected returns from business proposals

Unit A B C D
0 0 0 0 0
1 5 2 6 2
capital

2 6 4 7 3
3 7 6 8 4
4 8 8 8 5
5 8 9 8 6
6 8 10 8 6

3. a) State and explain Richard E. Bellman principal of optimality in dynamic

programming?
b) A manufacturing organization has the requirement of 100, 200, 300, 300 and 200
items for the ensuring five periods. The procurement cost is Rs.12 per procurement
irrespective of the size of the purchase. The holding cost is Rs.2 per 100 items per
period. The maximum inventory is not to exceed 400 items at any period. Formulate a
dynamic programming problem and find the policy of procurement so as to maximize
the total cost. No inventory is to be left in the last season maximum quantity ordered
is 600.
4. a)What are the essential characteristics of dynamic programming?
b) A manufacturing company has three sections producing automobile parts, computer
parts and electronic gadget parts respectively. The management has allocated Rs.20,
000 for expanding the production facilities in the automobile parts in the computer
parts section. The production can be increased either by adding new machines or by

29
Operations Research Dynamic programming

replacing some old inefficient machines by atomic machines. The electronic gadget
parts section was started only a few days back and thus additional amount can be
invested only by adding new machines to the section. The cost of adding and replacing
the machines along with the association expected returns in the different sections is
given in the table below. Select a set of expansion plans which may yield maximum
return. Use dynamic programming.
Automobile parts Computer parts Electronic gadgets
Alternatives Cost (Rs) Return Cost (Rs) Return Cost (Rs) Return
(Rs) (Rs) (Rs)
No 0 0 0 0 0 0
expansion
Add new 4000 8000 8000 12000 2000 8000
machines
Replace old 6000 10,000 12000 18000 - -
machines

4. Use dynamic programming to solve

Minimize Z=
Subject to: y1 + y2 + y3 = 5
y1, y2, y3

30