You are on page 1of 10

[OR PROJECT REPORT]

Submitted By

RESOURCE
Abhishek Mudiganti
Akankshit Kanungo
U110124
U110126
Ansuman Pradhan U110128
Siddharth Panda U110167

ALLOCATION
Sourav Dutta
Sumit Mishra
Section C Group 8
U110170
U110174

PROBLEM
OR Project Report Group 8 Sec C

A RESOURCE ALLOCATION PROBLEM


The federal government has agreed to provide funds for a project to transform some untrained and
unemployable people into occupationally trained employable individuals through a process of
training. If a participant completes the training period for a particular program, there is a positive
benefit as a result of the cost incurred in the training. If the participant does not complete the
training, there is no benefit, but a fractional cost of training is incurred anyway, depending on how
much of the period he or she remained in training. The problem is to allocate the funds and
participants among the various programs to maximize the total expected benefit, subject to the
budget and other constraints. Available data is tabulated at the bottom of this page (some of the
values are estimated from input information for a community of about 700,000 people consisting
primarily of disadvantaged persons, of whom a maximum of 500 are expected to apply for a training
program of 12 weeks duration). The total budget available is $500,000. The expected benefit is the
expected increased annual income earned by successful participants. Ignoring the integer
requirements on the xi, formulate the problem of determining the x, y to maximize the expected
total benefit as an LP.

Training Program
I=1 2 3 4
Description Work On-the-Job Classroom Public
Experience Training Training Service Job
ci = cost of program in dollars per participant 960 2400 2880 3600
accepted to take it
bi = the increased annual income in dollars 8320 10,400 10,400 7800
that a participant can expect to receive, if he
or she completes the program successfully
Φi = the probability that an applicant will 0.8 0.5 0.7 0.6
qualify to be accepted as a participant in the
program
Πi = the probability that a participant will 0.4 0.72 0.33 0.38
complete the program successfully if he or she
begins it
Θi = the probability that a client who 0.34 0.7 0.4 0.4
completes the program will succeed in
obtaining a job (giving him or her the expected
increase annual income)
ti = fractional cost incurred per participant 0.25 0.13 0.27 0.26
beginning in the program but not completing
it successfully
ni = maximum number of applicants that can 150 315 20 50
be screened by the program
xi = decision variable, the number of applicants X1 X2 X3 X4
sent to the ith program
yi = decision variables, dollars of budget Y1 Y2 Y3 Y4
allocated to the ith program
OR Project Report Group 8 Sec C

Solution:

Let X1, X2, X3 and X4 be the variables such that:

X1= No. of Applicants sent to the “Work Experience” Program

X2= No. of Applicants sent to the “On-the-Job Training” Program

X3= No. of Applicants sent to the “Classroom Training” Program

X4= No. of Applicants sent to the “Public Service Job” Program

The total number of candidates that can be screened by the programs are 150, 315, 20 and
50 respectively.

Therefore

X1 <= 150…………………………………………………………………………………………………………………………. (1)

X2 <= 315…………………………………………………………………………………………………………………………. (2)

X3 <= 20…………………………………………………………………………………………………………………………… (3)

and X4 <= 50……………………………………………………………………………………………………………………. (4)

Moreover, a maximum of 500 are expected to apply for the program.

So, X1 + X2 + X3 + X4 <= 500……………………………………………………………………………………………… (5)

The total Budget allocated for the entire program is $500,000. This has to be divided among
the training programs so that optimum allocation results.

Training Program 1:

Of all the applicants who have applied for Training 1 i.e. Work Experience Program, it is
expected that 80% will qualify to be the participants in the program.

If Applicants = X1,
Then Participants = 0.8*X1

Now from all the participants, 0.4 are expected to successfully complete the program. i.e.
The rest 0.6 won’t be able to complete the training program. Those who successfully
complete the training, incur a cost of $960 per person and those who don’t complete the
program, incur a fractional cost of 0.25*$960.

So,

Total cost incurred by Training Program 1 = 960*0.4*0.8*X 1 + 960*0.25*0.6*0.8*X1


OR Project Report Group 8 Sec C

= 422.4*X1

Training Program 2:

Similarly for Training Program 2, the total cost incurred

= 2400*0.72*0.5*X2 + 2400*0.13*0.28*0.5*X2

= 948*X2

Training Program 3:

Similarly for Training Program 3, the total cost incurred

= 2880*0.33*0.7*X3 + 2880*0.27*0.67*0.7*X3

= 1002.96*X3

Training Program 4:

Similarly for Training Program 4, the total cost incurred

= 2400*0.38*0.6*X4 + 2400*0.26*0.62*0.6*X4

= 1155.6*X4

Therefore,

422.4X1 + 948X2 + 1002.96X3 + 1155.6X4 <= 500000…………………………………………………………. (6)

And 422.4X1 = Y1, 948X2 = Y2, 1002.96X3 = Y3 and 1155.6X4 = Y4

Where Y1, Y2, Y3, Y4 are the costs incurred in program 1, 2, 3 and 4 respectively.

All X1, X2, X3, X4 >=0

Total Expected Benefit:

X1*0.8*0.4*0.34*8320 = 905.22*X1

X2*0.5*0.72*0.7*10400 = 2620.8*X2

X3*0.7*0.33*0.4*10400 = 960.96*X3

And X4*0.6*0.4*0.38*7800 = 711.36*X4

To maximize the expected benefit from the program,

Max Z = 905.22X1 + 2620.8X2 + 960.96X3 + 711.36X4


OR Project Report Group 8 Sec C

Thus, the LP formulation of the Resource Allocation Problem is given as,


Maximize Profit = 905.22X1 + 2620.8X2 + 960.96X3 + 711.36X4

Given,

X1 <= 150

X2 <= 315

X3 <= 20

X4 <= 50

X1 + X2 + X3 + X4 <= 500

422.4X1 + 948X2 + 1002.96X3 + 1155.6X4 <= 500000


OR Project Report Group 8 Sec C

CPLEX Code for the LP formulation:

Maximize
obj: 905.22x1 + 2620.8x2 + 960.96x3 + 711.36x4
Subject To
c1: x1 <= 150
c2: x2 <= 315
c3: x3 <= 20
c4: x4 <= 50
c5: x1 + x2 + x3 + x4 <= 500
c6: 422.4x1 + 948x2 + 1002.96x3 + 1155.6x4 <= 500000
Bounds
0 <= x1
0 <= x2
0 <= x3
0 <= x4
General
x1
x2
x3
x4
End
OR Project Report Group 8 Sec C

Solution Using CPLEX (Screen Shots):

(ScreenShot-1 , Feeding and optimizing the LP equations in CPLEX)

(Screen Shot-2, displaying solution of the variables (x1 to x4))


OR Project Report Group 8 Sec C

(Screen Shot-3, Display of slack variables)

(Screenshot-4, Populate the solution pool (21 solutions saved))


OR Project Report Group 8 Sec C

(Screenshot-5, Solution List (21 solutions))

(Screenshot-6, Display of Optimized Solution member 1)


OR Project Report Group 8 Sec C

Conclusion and Inferences:

Out of all the 21 solutions, the solution member 1(p1) got the highest objective value that
is maximum expected benefit of the program.

Solutions

Variable Solution member p1, with optimized


net profit
X1 (No. of applicants for Program 1) 150
X2 (No. of applicants for Program 2) 315
X3 (No. of applicants for Program 3) 20
X4 (No. of applicants for Program 4) 15
Y1 (Budget for program 1) $63360
Y2 (Budget for program 2) $298620
Y3 (Budget for program 3) $20059.20
Y4 (Budget for program 4) $17334

Inferences:

1. To achieve the maximum expected benefit, 150 applicants should be taken for “Work
Experience” program, 315 applicants should be taken for “On-the-Job Training” program, 20
applicants should be taken for “Classroom training” and 15 applicants should be taken for
“Public Service Job” program.
2. The maximum budget allocated for X1 should be $63360, for X2 should be $298620, for X3
should be $20059.20 and for X4 should be $17334.
3. Net profit for the program should be
Total increased benefit – Total cost incurred
=$991224.6 - $399373.2
=$591851.4

-------------------------------------------------------------Thank You-----------------------------------------------------------

You might also like