Symbiotic Organisms Search (SOS) algorithm based on B* tree Crossover for fixed outline VLSI floorplans

Abstract

In VLSI physical design vicinity, floorplanning is a crucial and powerful step for fixing circuit layout complexity that's getting extended because of a wider variety of additives are incorporated right into a single chip. Floorplanning provides a ground work to solve this problem by identifying the relative locations of modules(blocks) also estimates dead space (white space), total layout area (chip area) and wirelength among modules. This work describes a multi objective adaptive symbiotic organisms search (SOS) algorithm for soft modules with fixed outline. An adaptivity in Multi Objective Optimization (MOO) leads the way to metaheuristics for solving most of the real time problems be connected with electronics. A novel B*tree crossover operator is used by this SOS floorplanner. In traditional symbiotic organism’s search (SOS) optimization approach, crossover operation over B*tree is not attempted. The proposed SOS algorithm produces the effective combinations of B*tree structures as a result of crossover. Three phases of symbiotic organism’s search namely, mutualism, commensalism, and parasitism are effectively handled by this new B*tree structures. Search space exploration and exploitation are balanced by the effective combinations of B*tree structures. SOS algorithm results are compared with existing optimization methods mentioned in literature. The proposed SOS algorithm is more efficient in area and wirelength minimization than the state-of-the-art algorithms. MCNC (Microelectronics Center of North Carolina) benchmarks are used for testing SOS algorithm. Test results of SOS algorithm proves that better results are produced for wirelength minimization, area minimization and dead space minimization over previous floorplanning algorithms.