Development of realtime scheduler framework for task scheduling in multiprocessor environment

Abstract

The objective of this thesis is to develop the Multiprocessor newlineScheduling algorithm for real time application by considering the newlineoptimization techniques. The goal of the scheduling in optimization newlinetechniques is to find out the best processor amongst the group of processors. newlineHere, the finding of the best processor involves a search technique where the newlinetasks would be searching for the processor in a group. In this search criterion, newlinethe tasks remember the execution capability of each processor and goes to the newlinenext processor. It finally declares one processor as a best processor, which is newlinecalled a local best processor. In this way all the processors reveals their local newlinebest processor where one amongst them is decided as a global best processor. newlineA framework is developed in LABVIEW to illustrate the working newlineof developed algorithm with its parameters like velocity and position. newlineSimulation of multiprocessor system is an abstract representation of the newlineproposed or existing system as a software program. Simulation helps to newlineunderstand the behaviour of the multiprocessor system. newlineA hardware system is designed as a part of this research work to newlinedemonstrate the working of Particle Swarm Optimization, Cuckoo Search newlineAlgorithm, Fuzzy logic approximation and the Proposed Novel Algorithm. newlineThe hardware System consists of a water tank with four Light Dependent newlineResistors for Particle Swarm Optimization Algorithm, Ultrasonic Sensor for newlineCuckoo Search, three steel rods for measuring the height of the tank, and one newlinelight Dependent Resistor for Novel Algorithm. newlineIn the research work the parameters of the four algorithms Particle newlineSwarm Optimization, Cuckoo Search Algorithm, Fuzzy Logic Approximation newlineand Novel Algorithm are displayed in the window. newline