Date of Award


Publication Type

Master Thesis

Degree Name



Computer Science

First Advisor

Aggarwal, Akshaikumar (School of Computer Science)


Computer Science.




This thesis proposes a GA based scheduling algorithm for a heterogeneous distributed computing environment. It uses the application centric load balancing system. The proposed system removes all those network delay assumptions and considers the allocation problem for both computing resources and network resources. It addresses the generalized workload i.e. direct acyclic graph (DAG) kind of workload that is composed of sub jobs with internal dependencies. Rather than allocate applications simply according to their arrival time, we introduce GA scheduling strategy into our load balancing system to find the proper applications allocating schedule, which uses the available resources more efficiently. With introduction of GA scheduling into both application level and process level, certain improvements on the practicability, accuracy and performance are expected. Instead of using constant GA parameters, our proposed algorithm dynamically adjusts the key parameters, such as crossover rate and mutation rate, adapting them to the quality of generations. Later, we will implement more new ideas, such as gender assignment, fertility rate and aging into our GA algorithm to achieve better performance. 'Keywords.' Load balancing, network delay, workload simulator, Adaptive Genetic Algorithm scheduling etc.