Abstract:
This thesis describes the development and application of a decision support
system for logistics strategy modelling. The decision support system that is
developed enables the modelling of logistics systems at a strategic level for
any country or area in the world. The model runs on IBM PC or compatible
computers under DOS (disk operating system).
The decision support system uses colour graphics to represent the different
physical functions of a logistics system. The graphics of the system is machine
independent. The model displays on the screen the map of the area or country
which is being considered for logistic planning.
The decision support system is hybrid in term of algorithm. It employs
optimisation for allocation. The customers are allocated by building a network
path from customer to the source points taking into consideration all the
production and throughput constraints on factories, distribution depots and
transshipment points.
The system uses computer graphic visually interactive heuristics to find the
best possible location for distribution depots and transshipment points. In a
one depot system it gives the optimum solution but where more than one
depot is involved, the optimum solution is not guaranteed.
The developed model is a cost-driven model. It represents all the logistics
system costs in their proper form. Its solution very much depends on the
relationship between all the costs. The locations of depots and transshipment
points depend on the relationship between inbound and outbound
transportation costs.
The model has been validated on real world problems, some of which are
described here. The advantages of such a decision support system for the
formulation of a problem are discussed. Also discussed is the contribution of
such an approach at the validation and solution presentation stages.
