Designing and measuring the Performance of agent based Scheduling and load balancing for Multicore architecture

Abstract

The main objective of this research work is to develop an enhanced newlinescheduling and load balancing algorithm for multicore architecture In the newlinefirst phase of this research three different types of agent based scheduling newlinealgorithms are implemented and they increase the performance of the newlineprocessor by efficiently allocating the tasks to the processors In the first newlineproposed method the AMAS theory of multiagent system is combined with newlinethe scheduler of operating system to develop a new agent based scheduling newlinealgorithm for multicore architecture In the second proposed research method newlinehard soft processor affinity scheduling algorithm is implemented which newlineavoids the context switching of critical tasks In the third method a novel newlineagent based scheduling and thread assignment algorithm is proposed in such a newlineway that none of the heterogeneous processors will be kept in the idle state newlineand the cores are utilized efficiently newlineIn the second phase of the research simple load balancing algorithm newlineis proposed and it is a direct derivation and solution obtained from the newlinedefined agent scheduling algorithms newline newline