SCM 3301

Linear Programming Formulation of Max Flow Problem

The Maximal Flow Problem
• Objective:
• Determine the largest amount of flow that can enter or exit a network
• Limited by capacities on the individual arcs of the network
• Important applications:
•Transportation: management of flow on highway systems, urban traffic, etc.
•Pipeline (water, oil and gas, phosphate, etc.) system management
•Networking: bandwidth management, etc.
•Some applications in finance: management of stock exchange networks,
FOREX operations, etc.
•…
 Maximal Flow Problem – An Example
Proposed network for crossing the city of Glasgow, Scotland Arc Capacity = 8000 vehicles / hour from node 5 to node 7

i.e.
3
2 5 Arc (5,7) can accommodate any number of vehicles from 0
to 8000 per hour
8
5 2 3 1
2 1

6 7 7
1 3 6 7 Question:
5 5
Entering
3
Leaving
What is the maximum number of vehicles
Glasgow (W) Glasgow (E) (flow) that can travel from origin (node 1)
to destination (node 7) per hour?
4
 Maximal Flow Problem – An Example
Proposed network for crossing the city of Glasgow, Scotland

Decision variables:
3
2 5 = flow on each arc (i,j) in the network
8
5 2 3 1
2 1
In our case, we have:
6 7 7
1 3 6 7
= flow on arc (1,2)
5 5 = flow on arc (1,3)
3
Entering Leaving = flow on arc (1,4)
Glasgow (W) Glasgow (E)
= flow on arc (2,3)
4 …
 Maximal Flow Problem – An Example
Proposed network for crossing the city of Glasgow, Scotland

Objective:
3
2 5
Max++
8
5 2 3 1 or
2 1
Max+
6 7 7
1 3 6 7
5 5
3
Entering Leaving Maximize
Glasgow (W) Glasgow (E)
The total flow getting out of the origin (node 1)
4 or
The total flow getting in the destination (node 7)
 Maximal Flow Problem – An Example
Flow Out – Flow In = 0 for all nodes
Proposed network for crossing the city of Glasgow, Scotland except the origin and the destination

Constraints:
3
2 5 +-- =0
8 +++ -- =0
5 2 3 1
2 1 - =0
6 7 7 + --- =0
1 3 6 7 + --- =0

5 5
3
Entering Leaving ≤ 5, ≤ 6, ≤ 5, ≤ 2, ≤ 3,
Glasgow (W) Glasgow (E) …, ≤ 7
4
Capacity Constraints
 Maximal Flow Problem – An Example
Flow Out – Flow In = 0 for all nodes
including the origin and the destination
Proposed network for crossing the city of Glasgow, Scotland
Constraints:
3
2 5 ++- =0
8 +-- =0
5 2 3 1
2 1 +++ -- =0
7
- =0
6 7
1 3 6 7 + --- =0
+ --- =0
5 5 -- =0
3
Entering Leaving
Glasgow (W) Glasgow (E)
≤ 5, ≤ 6, ≤ 5, ≤ 2, ≤ 2,
4 …, ≤ 7

Capacity Constraints

We simplify the formulation by adding a

“dummy” artificial arc from the destination back
to the origin
New Objective: Max