Multiprocessor System Scheduling with Precedence and Resources Constraints by Ant Colony System

Abstract

An ant colony optimization (ACO) approach for the precedence and resource constrained multiprocessor scheduling problem is presented and evaluated. The algorithm employs distributed agents which mimic the way of ants to find a feasible solution to complete jobs. Traditionally, the multiprocessor scheduling problems only consider the precedence constraint and finding the minimum of maximum complete time. In ACO, a set of ant-like agents or software ants solve the problem under consideration through a cooperative effort. A two-dimensional grid assigns jobs on processors scheme is proposed. This scheme is a time dependent structure in scheduling the jobs. In the exploration of the search space, a delay solution generation rules and dynamic rules are proposed, which are not only capable to escape the local optimal solution, but can get the feedback by previously constructed solutions. Simulation results reveal that developed ant colony system algorithm ensures an appropriate approach of solving multiprocessor system scheduling problems with precedence and resources constrains.