Search-based job shop scheduling and sequencing: Extensions to the search sequencing rule

Abstract

The authors have previously presented a rule which calculates job priority as the weighted, linear sum of the basic quantities used in many other priority rules. The format of the rule enables a number of these other priority rules to be expressed within the general framework as specific choices of the coefficients. A process of search and simulation is used to determine the values of the coefficients on the basis of giving satisfactory performance for a particular manufacturing system. This paper further tests this approach in three key areas. First, the inclusion of a term which provides feedback of the state of the shop is examined. The results obtained show a slight improvement when queue length has been used as the measure of system state. Second, the use of different performance indices to measure the performance of the manufacturing system is tested. The approach is shown to give significantly better results than the use of simple priority rules. Third, the approach is tested on a variety of systems and the results tend to indicate that the approach can work on a wide range of problems.