Use of Integer Programming as a tool for improving
the distribution policies among actors of a demand and
supply chain
1.
Topic
ƒ
2.
Original
reference
3.
Keywords
4.
Description of
the problem
approached
5.
Methods and
Tools
6.
Results
G. Nicosia, G. Oriolo and A. Pacifici, "Optimal allocation plan
for distribution centers of a frozen food company", Triennial
Symposium on Transportation Analysis, TRISTAN IV, pp. 5559, June 2001, Portugal.
Distribution, Supply chain, Logistics, Integer Programming.
In this paper the distribution network of an Italian company
operating in the ice cream and frozen food industry is analysed.
The distribution network is organized as a multi agent network,
where agents cooperate in order to develop the best product
allocation policies in order to minimize the company costs. In
particular, the addressed problem is that of optimally allocating
products demand to distribution centres spread over the Italian
territory. This is done by developing a mixed integer programming
model. The computational experience compares the optimal
solution found with the actual distribution policies of the company
and show how to use the proposed model as a decision support
tool for the company management.
Mathematical programming tools and commercial software for
mixed integer programming
Development of models that can be used to derive the minimum
cost solution for the company and as a decision support tool for
the company management in order to decide which parts of the
demand and supply network can be improved.
www.polito.it – codesnet@polito.it
1
B – Information for the End Users and Links
7.1
What’s new ?
the mathematical programming model and approach adopted to
determine new product allocation policies among the agents of the
supply chain;
7.2
What’s useful ?
The proposed mathematical programming approach can be easily
modified and applied to other real life scenarios both for
minimizing costs and for deciding what changes might be
appropriate in the demand and supply network.
7.3
Where can be
used ?
It can be used with any company in which data on capacities,
demand, production, transportation, storage and handling costs
are available and in which the codesnet is a layered network.
8.
Tools
(technology)
for the
implementation
of results.
No specific tools are required. For the problem definition and
solution, standard mathematical programming techniques and
tools might be used. The implementation of the solution found,
since it regards product allocation among distribution centers,
does not require any specific technological tool.
9.
Links to
industrial cases
in the V-LAB
10.
Links to other
papers in the
V-LIB
11.
Notes
www.polito.it – codesnet@polito.it
2
C - Identification of the paper/report
Process
12.
Identification
Card
Network Design
Organization
Supply
Chain
20 %
Production
Service
Management
Logistic
Network
Development
Innovation
Policy
80 %
Structure
Industrial
District
D. Conceptual scheme of the paper contents:
DESNET Committee (DC) Chair
DESNET Agents
Scope:
improve distribution
system of a frozen
food company
Collects data on location of
logistic poles and distribution
flows.
Asks for data on costs,
demands, capacities, and
production.
All agents (Production Poles,
Primary Distribution Centers,
Secondary Distribution
Centers) send to the Chair
their data.
Builds and solves ILP model.
Either implements solution found or
studies a new scenario with different data.
www.polito.it – codesnet@polito.it
3
D (continue). Detailed description of the main paper result:
DESNET Committee (DC) Chair
1st Step:
Collect data or analyse performance of the system with the following data:
• location of logistic poles and distribution flows;
• costs and capacity information on Primary Distribution Centers;
• transportation costs;
• production data from Production Sites for each time period and product type;
• demand data for the Secondary Distribution Centers for each time period and product
type.
2st Step:
Build mixed integer programming model.
3rd Step:
Solve mixed integer program (by commercial software or ad hoc algorithm).
4th Step:
Analyse solution and possibly update data to evaluate a different situation.
www.polito.it – codesnet@polito.it
4