Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Operations Research I: Models & Applications

Linear Programming

Ling-Chieh Kung

Department of Information Management

National Taiwan University

OR I: Linear Programming 1 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Introduction

I Let’s study Linear Programming (LP).

I It is used a lot in practice.
I It also possesses useful mathematical properties.
I It is a good starting point for all OR subjects.
I We will study:
I What kind of practical problems may be solved by LP.
I How to formulate a problem as an LP.

OR I: Linear Programming 2 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Road map

I Terminology.
I The graphical approach.
I Three types of LPs.
I Simple LP formulations.
I Compact LP formulations.

OR I: Linear Programming 3 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Linear Programs

I Linear Programming is the process of formulating and solving linear

programs (also abbreviated as LPs).
I An LP is a mathematical program with some special properties.
I Let’s first introduce some concepts of mathematical programs.

OR I: Linear Programming 4 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Basic elements of a program

I In general, any mathematical program may be expressed as

min f (x1 , x2 , ..., xn ) (objective function)

s.t. gi (x1 , x2 , ..., xn ) ≤ bi ∀i = 1, ..., m (constraints)
xj ∈ R ∀j = 1, ..., n. (decision variable)
I There are m constraints and n variables.
I x1 , x2 , ..., and xn are real-valued decision variables.
I We may write
 
x1
x =  ...  = (x1 , ..., xn )
 

xn
as a vector of decision variables (or a decision vector).
I f : Rn → R and gi : Rn → R are all real-valued functions.
I Mostly we will omit xj ∈ R.

OR I: Linear Programming 5 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Transformation

I How about a maximization objective function?

I max f (x) ⇔ min −f (x).
I How about “=” or “≥” constraints?
I gi (x) ≥ bi ⇔ −gi (x) ≤ −bi .
I gi (x) = bi ⇔ gi (x) ≤ bi and gi (x) ≥ bi , i.e., −gi (x) ≤ −bi .
I For example:

max x1 − x2 min −x1 + x2

s.t. −2x1 + x2 ≥ −3 ⇔ s.t. 2x1 − x2 ≤ 3
x1 + 4x2 = 5. x1 + 4x2 ≤ 5
−x1 − 4x2 ≤ −5.

OR I: Linear Programming 6 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Sign constraints

I For some reasons that will be clear in the next week, we distinguish
between two kinds of constraints:
I Sign constraints: xi ≥ 0 or xi ≤ 0.
I Functional constraints: all others.
I For a variable xi :
I It is nonnegative if xi ≥ 0.
I It is nonpositive if xi ≤ 0.
I It is unrestricted in sign (urs.) or free if it has no sign constraint.

OR I: Linear Programming 7 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Feasible solutions

I For a mathematical program:

I A feasible solution satisfies all the constraints.
I An infeasible solution violates at least one constraint.
I For example:

min 2x1 + x2 I Feasible?

s.t. x1 ≤ 10 I x1 = (2, 3).
x1 + 2x2 ≤ 12 I x2 = (6, 0).
I x3 = (6, 6).
x1 − 2x2 ≥ −8
x1 ≥ 0
x2 ≥ 0.

OR I: Linear Programming 8 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Feasible region and optimal solutions

I The feasible region (or feasible set) is the set of feasible solutions.
I The feasible region may be empty.
I An optimal solution is a feasible solution that:
I Attains the largest objective value for a maximization problem.
I Attains the smallest objective value for a minimization problem.
I In short, no feasible solution is better than it.
I An optimal solution may not be unique.
I There may be multiple optimal solutions.
I There may be no optimal solution.

OR I: Linear Programming 9 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Binding constraints

I At a solution, a constraint may be binding:1

Definition 1
Let g(·) ≤ b be an inequality constraint and x̄ be a solution. g(·) ≤ b is
binding at x̄ if g(x̄) = b.

I An inequality is nonbinding at a point if it is strict at that point.

I An equality constraint is always binding at any feasible solution.
I Some examples:
I x1 + x2 ≤ 10 is binding at (x1 , x2 ) = (2, 8).
I 2x1 + x2 ≥ 6 is nonbinding at (x1 , x2 ) = (2, 8).
I x1 + 3x2 = 9 is binding at (x1 , x2 ) = (6, 1).

1 Binding/nonbinding constraints are also called active/inactive constraints.

OR I: Linear Programming 10 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Strict constraints?
I An inequality may be strict or weak:
I It is strict if the two sides cannot be equal. E.g., x1 + x2 > 5.
I It is weak if the two sides may be equal. E.g., x1 + x2 ≥ 5.
I A “practical” mathematical program’s inequalities are all weak.
I With strict inequalities, an optimal solution may not be attainable!
I What is an optimal solution of

min x
s.t. x > 0?

I Think about budget constraints.

I You want to spend $500 to buy several things.
I Typically, you cannot spend more than $500.
I But you may spend exactly $500.

OR I: Linear Programming 11 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Linear Programs

I A mathematical program I An example:

min f (x) min x1 + x2

s.t. gi (x) ≤ bi ∀i = 1, ..., m, s.t. x1 + 2x2 ≤ 6
2x1 + x2 ≤ 6
is an LP if f and gi s are all linear functions.
I Each of these linear functions may be expressed x1 ≥ 0, x2 ≥ 0.
as
n
X
a1 x1 + a2 x2 + · · · + an xn = aj xj ,
j=1

where aj ∈ R, j = 1, ..., n, are the coefficients.

I We may write a = (a1 , ..., an ) and f (x) = aT x.

OR I: Linear Programming 12 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Linear Programs

I In general, an LP may always be I Or by vectors:

expressed as
n min cT x
X
min cj xj s.t. aTi x ≤ bi ∀i = 1, ..., m.
j=1
Xn I ai ∈ Rn , bi ∈ R, c ∈ Rn .
s.t. Aij xj ≤ bi ∀i = 1, ..., m. I x ∈ Rn .
j=1 I Or by matrices:
I Aij s: constraint coefficients.
I bi s: right-hand-side values (RHS). min cT x
I cj s: objective coefficients. s.t. Ax ≤ b.

I A ∈ Rm×n , b ∈ Rm .

OR I: Linear Programming 13 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Road map

I Terminology.
I The graphical approach.
I Three types of LPs.
I Simple LP formulations.
I Compact LP formulations.

OR I: Linear Programming 14 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Graphical approach

I For LPs with only two decision variables, we may solve them with the
graphical approach.
I Consider the following example:

max 2x1 + x2
s.t. x1 ≤ 10
x1 + 2x2 ≤ 12
x1 − 2x2 ≥ −8
x1 ≥ 0
x2 ≥ 0.

OR I: Linear Programming 15 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Graphical approach

I Step 1: Draw the feasible region.

I Draw each constraint one by one, and then find the intersection.

max 2x1 + x2
s.t. x1 ≤ 10
x1 + 2x2 ≤ 12
x1 − 2x2 ≥ −8
x1 ≥ 0
x2 ≥ 0.

OR I: Linear Programming 16 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Graphical approach
I Step 2: Draw some isoquant lines.
I A line such that all points on it result in the same objective value.
I Also called isoprofit or isocost lines when it is appropriate.
I Also called indifference lines (curves) in Economics.

max 2x1 + x2
s.t. x1 ≤ 10
x1 + 2x2 ≤ 12
x1 − 2x2 ≥ −8
x1 ≥ 0
x2 ≥ 0.

OR I: Linear Programming 17 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Graphical approach
I Step 3: Indicate the direction to push the isoquant line.
I The direction that decreases/increases the objective value for a
minimization/maximization problem.

max 2x1 + x2
s.t. x1 ≤ 10
x1 + 2x2 ≤ 12
x1 − 2x2 ≥ −8
x1 ≥ 0
x2 ≥ 0.

OR I: Linear Programming 18 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Graphical approach
I Step 4: Push the isoquant line to the “end” of the feasible region.
I Stop when any further step makes all points on the isoquant line
infeasible.

OR I: Linear Programming 19 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Graphical approach

I Step 5: Identify the binding constraints at an optimal solution.

OR I: Linear Programming 20 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Graphical approach
I Step 6: Set the binding constraints to equalities and then solve the
linear system for an optimal solution.
I In the example, the binding constraints are x1 ≤ 10 and x1 + 2x2 ≤ 12.
I We may solve the linear system

x1 = 10
x1 + 2x2 = 12

in any way and obtain an optimal solution (x∗1 , x∗2 ) = (10, 1).
I For example, through Gaussian elimination:
     
1 0 10 1 0 10 1 0 10
→ →
1 2 12 0 2 2 0 1 1

I Step 7: Plug in an optimal solution obtained into the objective

function to get the associated objective value.
I In the example, 2x∗1 + x∗2 = 21.

OR I: Linear Programming 21 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Where to stop pushing?

I Where we push the isoquant line, where will be stop at?
I Intuitively, we always stop at a “corner” (or an edge).

I Is this intuition still true for LPs with more than two variables? Yes!
I A more rigorous definition of “corners” exists.

OR I: Linear Programming 22 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Road map

I Terminology.
I The graphical approach.
I Three types of LPs.
I Simple LP formulations.
I Compact LP formulations.

OR I: Linear Programming 23 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Three types of LPs

I For any LPs, it must be one of the following:

I Infeasible.
I Unbounded.
I Finitely optimal (having an optimal solution).
I A finitely optimal LP may have:
I A unique optimal solution.
I Multiple optimal solutions.

OR I: Linear Programming 24 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Infeasibility
I An LP is infeasible if its feasible region is empty.

min 3x1 + x2
s.t. x1 + x2 ≤ 4
3x1 + x2 ≥ 9
x1 − x2 ≤ 0.

OR I: Linear Programming 25 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Unboundedness
I An LP is unbounded if for any feasible solution, there is another
feasible solution that is better.

max x1 + x2
s.t. x1 + 2x2 ≥ 6
2x1 + x2 ≥ 6.

OR I: Linear Programming 26 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Unboundedness
I Note that an unbounded feasible region does not imply an
unbounded LP!
I Is it necessary?

min x1 + x2
s.t. x1 + 2x2 ≥ 6
2x1 + x2 ≥ 6.

I If an LP is neither infeasible nor unbounded, it is finitely optimal.

OR I: Linear Programming 27 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Multiple optimal solutions

I A linear program may have multiple optimal solutions.

min x1 + 2x2
s.t. x1 + 2x2 ≥ 6
2x1 + x2 ≥ 6
x2 ≥ 0.

I If the slope of the isoquant line is identical to that of one constraint,

will we always have multiple optimal solutions?

OR I: Linear Programming 28 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Summary

I In solving an LP (or any mathematical program) in practice, we only

want to find an optimal solution, not all.
I All we want is to make an optimal decision.

OR I: Linear Programming 29 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Road map

I Terminology.
I The graphical approach.
I Three types of LPs.
I Simple LP formulations.
I Compact LP formulations.

OR I: Linear Programming 30 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Introduction

I It is important to learn how to model a practical situation as an LP.

I Once you do so, you have “solved” the problem.
I This process is typically called LP formulation or modeling.
I Here we will give you some examples of LP formulation.
I Practice makes perfect!
I Then we formulate large-scale problems with compact formulations.

OR I: Linear Programming 31 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

A product mix problem

I We produce several products to sell.

I Each product requires some resources. Resources are limited.
I We want to maximize the total sales revenue with available resources.

OR I: Linear Programming 32 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Problem description

I We produce desks and tables.

I Producing a desk requires three units of wood, one hour of labor, and 50
minutes of machine time.
I Producing a table requires five units of wood, two hours of labor, and 20
minutes of machine time.
I We may sell everything we produce.
I For each day, we have
I Two hundred workers that each works for eight hours.
I Fifty machines that each runs for sixteen hours.
I A supply of 3600 units of wood.
I Desks and tables are sold at $700 and $900 per unit, respectively.

OR I: Linear Programming 33 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Define variables

I What do we need to decide?

I Let

x1 = number of desks produced in a day and

x2 = number of tables produced in a day.
I With these variables, we now try to express how much we will earn
and how many resources we will consume.

OR I: Linear Programming 34 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Formulate the objective function

I We want to maximize the total sales revenue.

I Given our variables x1 and x2 , the sales revenue is 700x1 + 900x2 .
I The objective function is thus

max 700x1 + 900x2 .

OR I: Linear Programming 35 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Formulate constraints
I For each restriction or limitation, we write a constraint:
Consumption per
Resource Total supply
Desk Table
Wood 3 units 5 units 3600 units
Labor 200 workers × 8 hr/worker
1 hour 2 hours
hour = 1600 hours
Machine 50 machines × 16 hr/machine
50 minutes 20 minutes
time = 800 hours
I The supply of wood is limited: 3x1 + 5x2 ≤ 3600.
I The number of labor hours is limited: x1 + 2x2 ≤ 1600.
I The amount of machine time is limited: 50x1 + 20x2 ≤ 48000.
I Use the same unit of measurement!

OR I: Linear Programming 36 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Complete formulation

I Collectively, our formulation is

max 700x1 + 900x2

s.t. 3x1 + 5x2 ≤ 3600 (wood)
x1 + 2x2 ≤ 1600 (labor)
50x1 + 20x2 ≤ 48000 (machine)
x1 ≥ 0
x2 ≥ 0.

I In any case:
I Clearly define decision variables in front of your formulation.
I Write comments after the objective function and constraints.

OR I: Linear Programming 37 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Solve and interpret

I An optimal solution of this LP is (884.21, 189.47).

I So the interpretation is... to produce 884.21 desks and 189.47 tables?
I “Producing 884.21 desks and 189.47 tables” seems weird, but in fact:
I We may produce 884.21 desks and 189.47 tables per day in average
(i.e., roughly 88,420 desks and 18,947 tables per 100 days).
I We may suggest to produce, e.g., 884 desks and 189 tables.2
I It still supports our decision making.
I It may not really be optimal, but we spend a very short time to make a
good suggestion.
I “All models are wrong, but some are useful.”

2 Why not 885 desks and 190 tables or the other two ways of rounding?
OR I: Linear Programming 38 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Produce and store!

I When we are making decisions, we may also consider what will happen
in the future.
I This creates multi-period problems.
I In many cases, products produced today may be stored and then sold
in the future.
I Maybe daily capacity is not enough.
I Maybe production is cheaper today.
I Maybe the price is higher in the future.
I So the production decision must be jointly considered with the
inventory decision.

OR I: Linear Programming 39 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Problem description
I We produce and sell a product.
I For the coming four days, the marketing manager has promised to
fulfill the following amount of demands:
I Days 1, 2, 3, and 4: 100, 150, 200, and 170 units, respectively.
I The unit production costs are different for different days:
I Days 1, 2, 3, and 4: $9, $12, $10, and $12 per unit, respectively.
I The prices are all fixed. So maximizing profits is the same as
minimizing costs.
I We may store a product and sell it later.
I The inventory cost is $1 per unit per day.3
I E.g., producing 620 units on day 1 to fulfill all demands costs

$9 × 620 + $1 × 150 + $2 × 200 + $3 × 170 = $6, 640.

3 Where does this inventory cost come from?

OR I: Linear Programming 40 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Problem description: timing

I Timing:

I Beginning inventory + production − sales = ending inventory.

I Inventory costs are calculated according to ending inventory.

OR I: Linear Programming 41 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Variables and objective function

I Let

xt = production quantity of day t, t = 1, ..., 4.

yt = ending inventory of day t, t = 1, ..., 4.
I It is important to specify “ending”!
I The objective function is

min 9x1 + 12x2 + 10x3 + 12x4 + y1 + y2 + y3 + y4 .

OR I: Linear Programming 42 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Constraints
I We need to keep an eye on our inventory:

I Day 1: x1 − 100 = y1 .
I Day 2: y1 + x2 − 150 = y2 .
I Day 3: y2 + x3 − 200 = y3 .
I Day 4: y3 + x4 − 170 = y4 .
I These are typically called inventory balancing constraints.
I We also need to fulfill all demands at the moment of sales:
I x1 ≥ 100, y1 + x2 ≥ 150, y2 + x3 ≥ 200, and y3 + x4 ≥ 170.
I Also, production and inventory quantities cannot be negative.

OR I: Linear Programming 43 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

The complete formulation

I The complete formulation is

min 9x1 + 12x2 + 10x3 + 12x4 + y1 + y2 + y3 + y4

s.t. x1 − 100 = y1
y1 + x2 − 150 = y2
y2 + x3 − 200 = y3
y3 + x4 − 170 = y4
x1 ≥ 100
y1 + x2 ≥ 150
y2 + x3 ≥ 200
y3 + x4 ≥ 170
xt , yt ≥ 0 ∀t = 1, ..., 4.

OR I: Linear Programming 44 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Simplifying the formulation

I May we simplify the formulation?
I Inventory balancing and nonnegativity imply demand fulfillment!
I E.g., in day 1, x1 − 100 = y1 and y1 ≥ 0 means x1 ≥ 100.
I So the formulation may be simplified to

min 9x1 + 12x2 + 10x3 + 12x4 + y1 + y2 + y3 + y4

s.t. x1 − 100 = y1
y1 + x2 − 150 = y2
y2 + x3 − 200 = y3
y3 + x4 − 170 = y4
xt ≥ 0, yt ≥ 0 ∀t = 1, ..., 4.

I Identifying redundant constraints (removing them does not alter the

feasible region) helps reduce the complexity of a program.

OR I: Linear Programming 45 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Simplifying the formulation

I One may further argue that there is no need to have ending inventory
in period 4 (because it is costly but useless).
I So the formulation may be further simplified to

min 9x1 + 12x2 + 10x3 + 12x4 + y1 + y2 + y3

s.t. x1 − 100 = y1 , y1 + x2 − 150 = y2
y3 + x3 − 200 = y3 , y3 + x4 − 170 = 0
xt ≥ 0 ∀t = 1, ..., 4
yt ≥ 0 ∀t = 1, ..., 3.

I However, this is not always suggested (at this stage).

I It is not required because a solver will see this.
I It is too difficult if the instance scale is large.
I In summary, simplification is good but in most cases unnecessary.

OR I: Linear Programming 46 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Personnel scheduling

I We are scheduling employees in a department store.

I Each employee must work for five consecutive days and then take
rests for two consecutive days.
I The number of employees required for each day:

Mon Tue Wen Thu Fri Sat Sun

110 80 150 30 70 160 120

I There are seven shifts: Monday to Friday, Tuesday to Saturday, ...,

and Sunday to Thursday.
I We want to minimize the number of employees hired.

OR I: Linear Programming 47 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Personnel scheduling

I We may find a feasible solution easily.

I For example, we may assign 150 employees to work from Monday to
Friday and 160 to work from Saturday to Wednesday:

Mon Tue Wen Thu Fri Sat Sun

Demand 110 80 150 30 70 160 120
Shift 1 150 150 150 150 150
Shift 6 160 160 160 160 160
Total 310 310 310 150 150 160 160

I This solution is feasible but seems to be bad.

OR I: Linear Programming 48 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Decision variables and objective function

I Let Monday be day 1, Tuesday be day 2, etc.

I Let xi be the number of employees who starts to work from day i for
five consecutive days.
I xi is the number of employees assigned to shift i.
I The objective function is thus:

min x1 + x2 + x3 + x4 + x5 + x6 + x7 .

OR I: Linear Programming 49 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Constraints
I Demand fulfillment:
I 110 employees are needed on Monday:

x1 + x4 + x5 + x6 + x7 ≥ 110.

I 80 employees are needed on Tuesday:

x1 + x2 + x5 + x6 + x7 ≥ 80.

I 120 employees are needed on Sunday:

x3 + x4 + x5 + x6 + x7 ≥ 120.

I Nonnegativity constraints:

xi ≥ 0 ∀i = 1, ..., 7.

OR I: Linear Programming 50 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Complete formulation

I The complete formulation is

min x1 + x2 + x3 + x4 + x5 + x6 + x7
s.t. x1 + x4 + x5 + x6 + x7 ≥ 110
x1 + x2 + x5 + x6 + x7 ≥ 80
x1 + x2 + x3 + x6 + x7 ≥ 150
x1 + x2 + x3 + x4 + x7 ≥ 30
x1 + x2 + x3 + x4 + x5 ≥ 70
x2 + x3 + x4 + x5 + x6 ≥ 160
x3 + x4 + x5 + x6 + x7 ≥ 120
xi ≥ 0 ∀i = 1, ..., 7.

OR I: Linear Programming 51 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Road map

I Terminology.
I The graphical approach.
I Three types of LPs.
I Simple LP formulations.
I Compact LP formulations.

OR I: Linear Programming 52 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Compact formulations

I Most problem instances in practice are of large scales.

I The number of variables and constraints are huge.
I Many variables may be grouped together:
I E.g., xt = production quantity of day t, t = 1, ..., 4.
I Many constraints may be grouped together:
I E.g., xt ≥ 0 for all t = 1, ..., 4.
I In modeling large-scale instances, we use compact formulations to
enhance readability and efficiency.
I We use the following three instruments:
I Indices (i, j, k, ...).
I Summation ( ).
P
I For all (∀).

OR I: Linear Programming 53 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Compact objective function

I The production-inventory problem:
I We have several periods. In each period, we first produce and then sell.
I Unsold products become ending inventories.
I We want to minimize the total cost.
I Indices: Because things will repeat in each period, it is natural to
use an index for periods. Let t ∈ {1, ..., 4} be the index of periods.
I For the objective function:

min 9x1 + 12x2 + 10x3 + 12x4 + y1 + y2 + y3 + y4 ,

if we denote the unit production cost on day t as Ct , t = 1, ..., 4, we

may rewrite it as
X4
min (Ct xt + yt ).
t=1

OR I: Linear Programming 54 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Compacting the constraints

I The original constraints:
I x1 − 100 = y1 , y1 + x2 − 150 = y2 , y2 + x3 − 200 = y3 , y3 + x4 − 170 = y4 .
I Let’s denote the demand on day t as Dt , t = 1, ..., 4:
I For t = 2, ..., 4 : yt−1 + xt − Dt = yt .
I We cannot apply this to day 1 as y0 is undefined!
I To group the four constraints into one compact constraint, we add an
additional decision variable y0 :

yt = ending inventory of day t, t = 0, ..., 4.

I Then the set of inventory balancing constraints are written as

yt−1 + xt − Dt = yt ∀t = 1, ..., 4.

I Certainly we need to set up the initial inventory: y0 = 0.

OR I: Linear Programming 55 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

The complete compact formulation

I The compact formulation is
4
X
min (Ct xt + yt )
t=1
s.t. yt−1 + xt − Dt = yt ∀t = 1, ..., 4
y0 = 0
xt , yt ≥ 0 ∀t = 1, ..., 4.
I Do not forget those for-all statements! Without them, the formulation
is wrong.
I Nonnegativity constraints for multiple sets of variables may be combined
to save some “≥ 0”.
I One convention is to:
I Use lowercase letters for variables (e.g., xt ).
I Use uppercase letters for parameters (e.g., Ct ).

OR I: Linear Programming 56 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Parameter declaration

I When creating parameter sets, we write something like

denote Ct as the unit production cost on day t, t = 1, ..., 4.

I Do not need to specify values, even though we have those values.
I Need to specify the range through indices.
I Parameter declarations should be at the beginning of the formulation.
I Parameters and variables are different.
I Variables are those to be determined. We do not know their values
before we solve the model.
I Parameters are given with known values.
I Parameters are exogenous and variables are endogenous.

OR I: Linear Programming 57 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Compact formulation for product mix

I Consider the product mix problem.
I Let n be the number of products and m be the number of resources.
I Let j and i be the indices for products and resources, respectively.
I We denote the unit sales price of product j as Pj , supply limit of
resources i as Ri , and unit of resource i required for producing one unit
of product j as Aij , where i = 1, ..., m, j = 1, ..., n.
I Let xj be the production quantity for product j, j = 1, ..., n.
I The compact formulation is
n
X
max Pj x j
j=1
Xn
s.t. Aij xj ≤ Ri ∀i = 1, ..., m
j=1

xj ≥ 0 ∀j = 1, ..., n.

OR I: Linear Programming 58 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Compact formulation for product mix

I Alternatively, let’s define J = {1, ..., n} as the set of products and

I = {1, ..., m} be the set of resources.
I The compact formulation is
X
max Pj x j
j∈J
X
s.t. Aij xj ≤ Ri ∀i ∈ I
j∈J

xj ≥ 0 ∀j ∈ J.

OR I: Linear Programming 59 / 60
Terminology Graphical approach Three types of LPs Simple formulation Compact formulation

Problems vs. instances

I A problem is an abstract description of a task to be completed or a
question to be solved.
I When we express everything with symbols, we have a problem.
I An instance is a concrete specification of a problem.
I When we plug in concrete values into symbols, we obtain an instance.
I A compact formulation like I A numeric formulation like

max 700x1 + 900x2

X
max Pj x j
j∈J s.t. 3x1 + 5x2 ≤ 3600
x1 + 2x2 ≤ 1600
X
s.t. Aij xj ≤ Ri ∀i ∈ I
j∈J 50x1 + 20x2 ≤ 48000
xj ≥ 0 ∀j ∈ J x1 ≥ 0, x2 ≥ 0

describes a problem. specifies an instance.

OR I: Linear Programming 60 / 60