Development of Local Clustering Organization Applied to Job-shop Scheduling Problem

A new method has been introduced to solve combinatorial problems so-named “local clustering organization method (LCO)”. In applying LCO to the traveling salesman problem (TSP), it is proven that LCO has good solutions with high accuracy and computation speed. A learning theory of LCO is based on Riccati type learning equation as well as self-organizing maps (SOM). However, LCO is independent of neuron synapses in learning process. LCO makes use of a criteria function instead of neuron synapses. In this study, The LCO solution is proposed in being applied to the job-shop scheduling problem (JSP) and its efficiency is investigated. Numerical experiments verify that LCO solves JSP with high accuracy and computation speed in comparison with the genetic algorithm.