Genetic computing for the film copy deliverer problem

Abstract

This paper studies the possibility and efficiency of applying genetic algorithms to the film copy deliverer problem which is a new combinatorial network model proposed by M. Gen et al (1993) and reformulated by L. Zhang and W.M. Zheng (1994). The problem belonging to NP-hard class is of both general meaning in theory and rather transparent background in engineering. A permutation-based genetic algorithm is devised to compute heuristic (including optimal) solutions of the problem and it provides, essentially, a general architecture open to many genetic crossover and mutation operators. The experiment simulations have primarily shown that the architecture enables some crossover operators such as PMX, CX and OX to produce feasible solutions, and contained the comparison of those operators during which two ways of mutations matched to the crossovers are considered.