Paper Details: Downloads: 442
Serial Number: P1120805002
Title: A FAST MULTI-OBJECTVE GENETIC ALGORITHM FOR HARDWARE-SOFTWARE PARTITIONING IN EMBEDDED SYSTEM DESIGN
Authors: M.Jagadeeswari and M.C.Bhuvaneswari
Abstract: This paper proposes a novel Multi-Objective Evolutionary Algorithm for hardware software partitioning of embedded systems. Customized genetic algorithms (GA) have been effectively used for solving complex optimization problems (NP Hard) but are mainly applied to optimize a particular solution with respect to a single objective. Many real world problems in embedded systems have multiple objective functions like area, performance, power, latency etc., which are to be maximized or minimized at the early stage of the design process. Hardware- software partitioning of embedded systems involves partitioning the system specification into hardware and software implementations with the goal to find a set of implementations that satisfy a number of constraints on cost and performance. In this paper a novel multi-objective algorithm called elitist non-dominated sorting genetic algorithm (NSGA-II) is applied to search for multiple optimal solutions, the knowledge of which helps the designer to compare and choose a compromised optimal solution for which hardware/software design can be implemented. The algorithm was implemented in C programming language. The application and adaptation of the NSGA-II algorithm and Weighted-Sum genetic Algorithm (WSGA) was analyzed for a well known 8-point FFT algorithm which can also be extended for 16-point FFT etc. From the simulation results NSGA-II was found to perform better than WSGA.
Keywords: Hardware-Software Partitioning, Embedded Systems, Genetic algorithm, NSGA-II, Pareto-optimal solutions.
Journal/Conference: International Journal of Artificial Intelligence and Machine Learning
Volume: 8
Issue: 2
Submission Date: 2/1/2008 12:00:00 AM
Review Date: 6/1/2008 12:00:00 AM
Publishing Date: 9/1/2008 12:00:00 AM
Article Downloads: 442
Download:

Facebook