Task Scheduling with Parallel Genetic Environment using Stepping Stone Method

In multiprocessor system Task scheduling is essential operation. The main objective of task scheduling is to shorten the length of schedule. The effectiveness by doing so is beneficial for large number of calculations having some constraints like time constraints etc. The proposed algorithm has the efficient execution of the schedule on parallel system that takes the structure of the application and the performance characteristics. Number of appriximation and heuristics algorithms have been proposed to fulfill the task scheduling problem. It is well known NP-Hard problemt. Here the study proposes a genetic based techniques to schedule parallel tasks on hetrogeneous parallel system. In this paper the scheduling problem considered includes a new heuristic algorithm for task scheduling, based on evolutionary method which embeds a new fast technique named Stepping Stone into Genetic Algorithm (GA). By comparing the proposed algorithm with an existing GA based algorithm, it is found that the computation time of the new algorithm to find a sub-optimal schedule is decreased; however, the length of schedule or the finish time is decreased too.