IQDM

Session 2
Questions?
 Linear Functions

• Is f = xy + yz a linear function?

• f(x,z) = xy + yz?

• f(y) = xy + yz?

• f(x,y) = xy + yz?
LP or LPP Formulation, in general

• Decision variables

• Objective function

• Constraints
 LP Assumptions
• Proportionality
➢ E.g. z = 4x + … .
➢ That is, 1 unit increase in x results in 4 units increase in z.

• Additivity
➢ If the above z is also proportional to y, where 1 unit increase in y
results in 3 units increase in z as well, then
➢ z = 4x + 3y

• Certainty
➢ Coefficients are deterministic

• Variables can take any real number values

 Example 2
Company X, an electronics company, has
designed two specialty add-on circuit cards for
personal computers and markets them
directly to the public.

One is a communication card, and the other a

video card. Sales of the two cards do not
“cannibalize” one another.
Direct production costs are $300 per
communication card and $400 per video card.

Marketing department estimates that the

demand for communication cards will be
10,000 – 9p units, where p is the retail price in
dollars.
Video card demand is similarly estimated to be
12,000 – 12t units, where t represents the price.

Company X has a total production capacity of 7,500

cards per year.

Assuming the marketing estimates are correct, how

much of each kind of card should company X
produce to maximize this year’s contribution?
Decision variables:
• Prices p and t

Maximize total contribution

Subject to constraints
• Is this an LP?

• Objective function, total contribution =

(p − 300)(10000 – 9p) + (t − 400)(12000 – 12t)

• Anyone who has seen this problem?

• Example from “Note on Linear Programming”
 Example 3: Staffing Problem
The Clark County Sheriff’s department schedules police
officers for 8-hour shifts.

The beginning times for the shifts are 8AM, noon, 4PM,
8PM, midnight and 4AM.

An officer beginning a shift at any of these times works for

the next 8 hours.

During normal operations, the number of officers needed

varies depending on the time of the day.

Management Science/Operations Research: A Strategic Approach by David Bell

 The department staffing guidelines require the following
minimum number of officers on duty.

Time of Day Minimum officers on Duty

8:00AM – Noon 5
Noon – 4:00PM 6
4:00PM – 8:00PM 10
8:00PM – Midnight 7
Midnight – 4:00AM 4
4:00AM – 8:00AM 6

The department wants to minimize the total number of officers

needed to meet all shift requirements. Formulate as an LP.
Decision Variables
x1= Number of officers that begin work at 8AM & work shifts (8AM - Noon) &
(Noon to 4PM)
x2= Number of officers that begin work at Noon & work shifts (Noon - 4PM) &
(4PM - 8PM)
x3= Number of officers that begin work at 4PM & work shifts (4PM – 8PM) & (8PM
to Midnight)
x4= Number of officers that begin work at 8PM & work shifts (8PM - Midnight) &
(Midnight – 4AM)
x5= Number of officers that begin work at Midnight & work shifts (Midnight –
4AM) & (4AM to 8AM)
x6= Number of officers that begin work at 4AM & work shifts (4AM – 8AM) &
(8AM - Noon)
Objective Function

Minimize the total number of officers needed to meet all shift

requirements.

That is:
min z = x1 + x2 + x3 + x4 + x5 + x6
 Constraints

Each shift must have at least the minimum

number of officers
Time of Day Minimum officers on Duty

8:00AM – Noon 5
Noon – 4:00PM 6
4:00PM – 8:00PM 10
8:00PM – Midnight 7
Midnight – 4:00AM 4
4:00AM – 8:00AM 6
 Constraints
Each shift must have at least the minimum number of
officers
• x6+x1 ≥ 5
• x1+x2 ≥ 6
• x2+x3 ≥ 10
• x3+x4 ≥ 7
• x4+x5 ≥ 4
• x5+x6 ≥ 6

Non-negativity constraints:
x1 ≥ 0, x2 ≥ 0, x3 ≥ 0, x4 ≥ 0, x5 ≥ 0, x6 ≥ 0
Integer constraints:
x1 , x2 , x3 , x4 , x5 and x6 are integers
 LP Formulation
DVs: x1 , x2 , x3 , x4 , x5 , x6
= The number of officers who begin work at 8AM, Noon,
4PM, 8PM, Midnight and 4AM, respectively.

min z = x1 + x2 + x3 + x4 + x5 + x6
s.t.
x6+x1 ≥ 5
x1+x2 ≥ 6
x2+x3 ≥ 10
x3+x4 ≥ 7
x4+x5 ≥ 4
x5+x6 ≥ 6
x1 ≥ 0, x2 ≥ 0, x3 ≥ 0, x4 ≥ 0, x5 ≥ 0, x6 ≥ 0; all xi int
 Example 4: Investment Problem

A portfolio manager in charge of a bank

portfolio has $10 million to invest.

The securities available for purchase, as well as

their respective quality ratings, maturities, and
yields, are shown in the following table:

Bradley, Hax, Magnanti

Bond name Bond type Quality Years to Yield to After-tax
maturity maturity yield
A Municipal 2 9 4.3 4.3
B Agency 2 15 5.4 2.7
C Government 1 4 5.0 2.5
D Government 1 3 4.4 2.2
E Municipal 5 2 4.5 4.5
The bank places the following policy limitations
on the portfolio manager’s actions:
1. Government and agency bonds must total at
least $4 million.
2. The average quality of the portfolio cannot
exceed 1.4 on the bank’s quality scale. (Note
that a low number on this scale means a
high-quality bond.)
3. The average years to maturity of the portfolio
must not exceed 5 years.
If you are the portfolio manager, what would
your objective be?

What all factors limit your objective?

(What are the constraints?)
 Decision Variables
Let xA , xB , xC , xD and xE

be the amounts (in millions of dollars)

to be invested in bonds A, B, C, D and E,

respectively.
 Objective
• Maximize the total after tax returns.

• That is:

max z =
0.043 xA + 0.027 xB + 0.025 xC + 0.022 xD + 0.045 xE
Bond name Bond type Quality Years to Yield to After-tax
maturity maturity yield
A Municipal 2 9 4.3 4.3
B Agency 2 15 5.4 2.7
C Government 1 4 5.0 2.5
D Government 1 3 4.4 2.2
E Municipal 5 2 4.5 4.5
 Constraints
1. Government and agency bonds must total at
least $4 million.

2. The average quality of the portfolio cannot

exceed 1.4 on the bank’s quality scale.

3. The average years to maturity of the portfolio

must not exceed 5 years.
Government and agency bonds must
total at least $4 million

xB + xC + x D  4
Bond name Bond type Quality Years to Yield to After-tax
maturity maturity yield
A Municipal 2 9 4.3 4.3
B Agency 2 15 5.4 2.7
C Government 1 4 5.0 2.5
D Government 1 3 4.4 2.2
E Municipal 5 2 4.5 4.5
 The average quality of the portfolio
cannot exceed 1.4
2 xA + 2 xB + xC + xD + 5 xE
gives the “total” quality (points).

Average is
(2xA + 2xB + xC + xD + 5xE)/(xA + xB + xC + xD + xE ).

The constraint is “Average  1.4”.

That is,
2xA + 2xB + xC + xD + 5xE  1.4(xA + xB + xC + xD + xE )
 The average years to maturity of the
portfolio must not exceed 5 years

(9xA + 15xB + 4xC + 3xD + 2xE)/

(xA + xB + xC + xD + xE)

5
Anything else?
xA + xB + x C + xD + xE

 10
 Non-negativity

xA , xB , xC , xD , xE  0
 Example 5
• A product is assembled with 4 units of component A
and 3 units of component B.

• The manufacturing shop runs three different

processes, each of which requires varying amounts
of raw materials and produce different amounts of
A and B.

• Two types of raw materials are used. 100 units of

raw material I (RM I) and 200 units of raw material II
(RM II) are available to the shop each day.
TTN, Proc. AWTOR
The following table gives the information on the
quantities of raw materials consumed by each
process and the yield of A and B from each of
them.

Input per run Output per run

Shop (units) (units)

RM I RM II A B
I 7 5 6 4
II 4 8 5 8
III 2 7 7 3
Formulate as an LP to maximize the number of
completed assemblies produced each day.
 Decision Variables

Let x1, x2 and x3 be the number of runs of

processes I, II and III, respectively.
Constraints
100 units of raw material I (RM I) and 200 units
of raw material II (RM II) are available to the
shop each day.

Input per run Output per run

Shop (units) (units)

RM I RM II A B
I 7 5 6 4
II 4 8 5 8
III 2 7 7 3
 Constraints

• 7x1 + 4x2 + 2x3 ≤ 100

• 5x1 + 8x2 + 7x3 ≤ 200

• x1, x2, x3 ≥ 0
Objective Function?
100 units of raw material I (RM I) and 200 units
of raw material II (RM II) are available to the
shop each day.

Input per run Output per run

Shop (units) (units)

RM I RM II A B
I 7 5 6 4
II 4 8 5 8
III 2 7 7 3
 “A product is
assembled with
4 units of
• Total quantity of A produced component A
= 6x1 + 5x2 + 7x3 and 3 units of
component B”.

• Total quantity of B produced

= 4x1 + 8x2 + 3x3

• Number of completed assemblies

= min { (6x1 + 5x2 + 7x3) / 4 ,
(4x1 + 8x2 + 3x3) / 3 }
 Objective
Maximize min { (6x1 + 5x2 + 7x3) / 4 ,
(4x1 + 8x2 + 3x3) / 3 }

Linear?

Homework: Linearize the above.

(That is, write linearly).
 Homework
A senior executive of a company recently opted for
VRS with a hefty packet of Rs. 100 crores. A Chit
Fund Company has offered the following
investment scheme for the benefit of such retired
people:
• "Invest a certain sum at the beginning of any
month, invest half of that amount beginning of
the next month and end of the second month,
you will get twice the amount invested originally
in the first month".

TTN, Proc. AWTOR

• This scheme is available for the next six months.

• The returns received at the end of any month can

be used immediately for reinvesting either as a
fresh investment or as a follow-up investment.

If you are in his/ her position, what will your aim be?

How will you achieve your aim/ objective?

Questions?
Thank You