Ant Colony and SVM

Ant Colony and SVM

Part I

Research Problem

Efficient task scheduling to various processors for any application is important to attain high performances. People look for alternative approaches in order to solve the problem, even if there are various conventional approaches available. The study which is based on scheduling of tasks uses Ant Colony Optimization to arrive at schedule.

Research Methodology

Characteristics of Processor

It is assumed that the processors are varied, in this view, the heterogeneity of the processors is based on the mixed proportional utilization of the similar task on different processors.

Characteristics of Tasks

It is assumed that the tasks are independent; therefore, there are no inter task communication and no precedence constraints among them. The utilization of a processor by a task is known a priori and it does not change with time. All the tasks are assumed to arrive at the same instant at 0 time units.

Results

The task scheduling algorithm based on Ant Colony Optimization is implemented, which is run for eight problem cases with the eight number of tasks and eight number of processors. The number of ants used for Ant Colony Optimization is 100. The results of the study presented that the individual utilization of processor obtained for 10 trial runs for task sets of size 150 and 80 respectively. It is found that through the computation of the average wait time of utilization of processors and tasks through the First Come First Served; the results presented that in view of the wait time Ant Colony Optimization, the First Come First Served performed better. However, the utilization of processors is more for a small number of processors through the First Come First Served algorithm; it is noted that among the processors in Ant Colony Optimization, the load is better.

Summary of the Article

The study relates to scheduling of ...