Improving production systems of timber-processing plants IV: Application of genetic algorithms for appropriate positioning of processing machines

The development of a layout plan for a new plant with the aid of genetic algorithms was studied to place the machines so that the plant floor was effectively utilized and the operation would not be impeded. Genetic algorithms are search algorithms based on the mechanics of natural evolution and natural genetics to solve problems in engineering fields. Simulation with the aid of genetic algorithms was undertaken step by step. The first seven hundred strings (chromosomes) were generated at random to organize an initial population. Each string consisted of 40 bits (genes), which represented characteristics of machines (x- andy-coordinates and inlet and outlet formations of materials on machines) in the binary coding. Then the simulation was undertaken by repeating selection, crossover, reproduction, and mutation of strings until all strings were saturated with the highest evaluation (fitness of chromosomes to environments in the case of creatures). Under some limitations, an acceptable layout plan of the modeled plant involving four wood-processing machines was obtained according to evaluation indices.Part III of this series appeared inMokuzai Gakkaishi 37: 702–710, 1991.Part of this report was presented at the 47^th annual meeting of the Japan Wood Research Society, Kocht, April 1997