Optimization of Simulated Annealing Parameters for Bi-Criteria Multiprocessor Task Scheduling Using RSM-Grey Methodology

Abstract

Abstract

Simulated Annealing (SA) has been shown as an exceptional method for finding an optimal solution to the combinatorial optimisation problems. In the present work, optimal combinations of different simulated annealing parameters have been identified by using the RSM-based grey relational analysis (GRA) for Bi-criteria multiprocessor task scheduling problem. Parameter selection is needed as SA is a meta-heuristic that does not specify all the necessary information for the optimised results. The experiments include 135 sets of data corresponding to samplings obtained from Response surface methodology (RSM) small factorial experimental design for three numeric parameters i.e., initial temperature, Re-anneal interval and weight of objective function with two categorical factors i.e., temperature function and move function, respectively. The multiprocessor task scheduling problem of standard LU decomposition with 14 tasks and 4 processors has been considered for simulated annealing parameters optimization. The objective function considered is minimization of the weighted sum of makespan and total completion time simultaneously.

Keywords: Multiprocessor task scheduling, simulated annealing, response surface methodology, makespan, total completion time

Cite this Article

Sunita Dhingra. Optimization of Simulated Annealing Parameters for Bi-Criteria Multiprocessor Task Scheduling Using RSM-Grey Methodology. Journal of Advancements in Robotics. 2017; 4(3):  17–28p.