Optimal Solution for VLSI Physical Design Automation Using Hybrid Genetic Algorithm

• Published in: Mathematical problems in engineering Vol. 2014; no. 1
• Main Authors: Shanavas, I. Hameem, Gnanamurthy, R. K.
• Format: Journal Article
• Language: English
• Published: New York Hindawi Publishing Corporation 01.01.2014; John Wiley & Sons, Inc
• Subjects: Automation; Benchmarks; Chemical partition; Codes; Design; Design optimization; Evolutionary algorithms; Genetic algorithms; Heuristic; Interconnect; Interconnections; Large scale integration; Mathematical analysis; Minimization; Objectives; Optimization; Optimization techniques; Partitioning; Placement; Simulated annealing; Very large scale integration
• ISSN: 1024-123X; 1026-7077; 1563-5147; 1563-5147
• DOI: 10.1155/2014/809642

In Optimization of VLSI Physical Design, area minimization and interconnect length minimization is an important objective in physical design automation of very large scale integration chips. The objective of minimizing the area and interconnect length would scale down the size of integrated chips. To meet the above objective, it is necessary to find an optimal solution for physical design components like partitioning, floorplanning, placement, and routing. This work helps to perform the optimization of the benchmark circuits with the above said components of physical design using hierarchical approach of evolutionary algorithms. The goal of minimizing the delay in partitioning, minimizing the silicon area in floorplanning, minimizing the layout area in placement, minimizing the wirelength in routing has indefinite influence on other criteria like power, clock, speed, cost, and so forth. Hybrid evolutionary algorithm is applied on each of its phases to achieve the objective. Because evolutionary algorithm that includes one or many local search steps within its evolutionary cycles to obtain the minimization of area and interconnect length. This approach combines a hierarchical design like genetic algorithm and simulated annealing to attain the objective. This hybrid approach can quickly produce optimal solutions for the popular benchmarks.