Discrete Multi Objective Particle Swarm Optimization Algorithm for FPGAPlacement

Placement process is one of the vital stages in physical design. In this stage, modules and elements ofthe circuit are placed in distinct locations based on optimizationprocesses. Hence, each placementprocess influences one or more optimization factor. On the other hand, it can be statedunequivocallythat FPGA is one of the most important and applicable devices in our electronic world. So, it is vital tospend time forbetter learning of its structure. VLSI science looks for new techniques for minimizingthe expense of FPGA in order to gain betterperformance. Diverse algorithms are used for runningFPGA placement procedures. It is known that particle swarm optimization (PSO) is one of the practicalevolutionary algorithms for this kind of applications. So, this algorithm is used for solving placementproblems. In this work, a novel method for optimized FPGA placement has been used. According tothis process, the goal is to optimize two objectives defined as wire length and overlap removalfunctions. Consequently, we areforced to use multi-objective particle swarm optimization (MOPSO)in the algorithm. Structure of MOPSO is such that it introduces set of answers among which we havetried to find a unique answer with minimum overlap. Itis worth noting that discrete nature of FPGAblocks forced us to use a discrete version of PSO. In fact, we need a combination of multi-objectivePSO and discrete PSO for achieving our goals in optimization process. Tested results on some ofFPGA benchmark (MCNC benchmark) are shown in “experimental results” section, compared withpopular method “VPR”. These results show that proper selection of FPGA’s size and reasonablenumber of blocks can giveus good response.