A genetic approach for solving a scheduling problem in a robotized analytical system

We consider a robotized analytical system in which a chemical treatment has to be performed on a given set of identical samples. The objective is to carry out the chemical treatment on the whole set of samples in the shortest possible time. All constraints have to be satisfied since a modification of the chemical process could create unexpected reactions. We have developed a new robust method governed by a genetic algorithm to solve this scheduling problem. The crossover mechanism of this evolutionary method is based on an extension of the uniform crossover introduced by Syswerda. The proposed approach can be adapted to other combinatorial problems where decisions, based on rules, have to be taken at each step of a constructive method.