FLLBHGATS: Efficient Load Balancing and Task Scheduling Algorithm for Real-Time Multiprocessor

Main Article Content

Nirmala H., Girijamma H. A.

Abstract

Different experiment has been advertised that the processor work load distributing equitably with the processors of a distributed system decidedly enhance framework execution and improves system management. Fuzzy logic has been implemented in numerous areas of industry and science to manage susceptibility. Proposed work with the intent of load balancing has been focused on using fuzzy logic to interpret processor’s load and task execution length. This work introduces a new dynamic fuzzy-based load balancing algorithm for homogeneous dispersed frameworks. The proposed techniques use fuzzy logic to manage improper data load i.e., overloaded and under loaded, deciding on load distribution choices and preserve general framework strength. For accurately evaluating the load status of a host, proposed algorithm uses CPU utilization, CPU queue length and distance upon its present load as linguistic inputs while framing fuzzy set. Method proposes Hybrid Genetic Algorithm (HGA) that is blended with stochastic development process in order to designate and schedule real-time tasks with priority requirements. The work randomly generates the tasks using random wheel approach, once the tasks are generated then encoding tasks to chromosome is carried out. Height of each task is obtained through DAG and according to the root node, the height of each taskis updated in the chromosome. Proposed fuzzylogicbased load balancing and hybrid genetic algorithm based task scheduling (FLLBHGATS) algorithm has been evaluated with similar existing methods in order to prove its efficiency. The results prove that FLLBHGATS performs better than other techniques as far as the solution quality.

Article Details

Section
Articles