Viewing scheduling problems through genetic and evolutionary algorithms

Abstract

In every system, where the resources to be allocated to a given set of tasks are limited, one is faced with scheduling problems, that heavily constrain the enterprise's productivity. The scheduling tasks are typically very complex, and although there has been a growing flow of work in the area, the solutions are not yet at the desired level of quality and efficiency. The Genetic and Evolutionary Algorithms (GEAs) offer, in this scenario, a promising approach to problem solving, considering the good results obtained so far in complex combinatorial optimization problems. The goal of this work is, therefore, to apply GEAs to the scheduling processes, giving a special attention to indirect representations of the data. One will consider the case of the Job Shop Scheduling Problem, the most challenging and common in industrial environments. A specific application, developed for a Small and Medium Enterprise, the Tipografia Tadinense, Lda, will be presented.