A Self-Adaptive Multi-Resources Suppliers System

Abstract

Dealing with a suppliers/demands system with functional transportation cost is usually considered as a classical transportation problem and can be solved with normal linear programming processes, but what adds to it, when the cost in the supply points is to be taken into consideration, demands at each demand point varies with time, and complexity of the whole system grows. Furthermore, introducing a constant cost to services and/or transportations, which is the case in several real distributed systems, makes the model more sophisticated to be solved. In this case, additional measures prevent us from treating the model as a classical transportation problem; instead, we have to search another technique to grant satisfactory results. This paper proposes an approach that enables a system to autoreconfigure itself at runtime whenever environmental variables vary, with the goal of minimizing costs while satisfying the demands, while taking into consideration the different type of costs (constant/functional, service/transportation, etc.). Furthermore, this work aims to build a framework to find the optimal configuration of services/transportations in a multiresources suppliers system requiring an auto adaptation facing the variation of demands. Our proposed framework, called GATRA, is a combination of the Genetic Algorithms and the Transportation problem techniques.