A genetic algorithm based approach for multi objective optimization of real time scheduling

Abstract

With the many fold increase in CPU processing power the use of newlinespinning magnetic disks for data storage is very crucial For real time newlineapplications like multimedia the order in which the read write requests are newlinesubmitted to storage volumes is very important Arranging the readwrite newlinerequests is called disk scheduling Since the invention of movable head disks newlineseveral algorithms have been developed to improve the performance of disk newlinescheduling Genetic algorithm based approaches are proposed in this thesis to newlineoptimize disk scheduling The crucial challenge that decides the success of newlinerealtime disk scheduling algorithm lies in simultaneously achieving the two newlinecontradicting objectives namely completion time and missed tasks This newlinework is motivated toward developing such an algorithm The goal of this newlinework is to demonstrate that the simultaneous optimization of completion time newlineand missed tasks produces an efficient schedule for realtime disk scheduling newlineThe objective function is designed to minimize the two parameters newlinecompletion time and the number of missed tasks newline newline