Performance Evaluation of Simulated Annealing-Based Task Scheduling Algorithms

Abstract

The performance of simulated annealing (SA)-based task scheduling algorithms is evaluated. First, various parameters of SA are varied, and it is seen how it affects the schedule length (SL). The parameters that are varied are initial temperature, number of iterations, initial clustering, and cooling schedule. Then, one SA-based task scheduling algorithm is selected and compared with other task scheduling algorithms. The algorithms selected for comparison are cluster pair priority scheduling (CPPS), dominant sequence clustering (DSC), edge zeroing (EZ), and linear clustering (LC). Random task graphs are used for comparison