Development of a Predictive -reqactive Scheduler Using Genetic Algorithms and Simulation -based Scheduling Software Property of Msld Library

SYNOPSIS Simulation-based scheduling packages are widely used in manufacturing plants around the world. These packages include a large number of " canned " dispatching rules which produce schedules for both simple and complex production environments. Users can extend these canned rules by adding their own plant-specific rules. In this paper we describe an iterative approach in which we first generate schedules using an optimal search technique, genetic algorithms (GA), and then predict how well those schedules perform using the simulation-based scheduling software, We also describe an information model for a status database which provides the basis for using this iterative approach for reactive scheduling as well.