An Adaptive Hybrid Genetic Algorithm for VLSI Standard Cell Placement Problem

This paper presents an adaptive hybrid genetic algorithm (AHGA) for VLSI standard cell placement problem which belongs to NP-hard combinatorial optimization problem. Based on the distinguishing feature of solution space of the problems with various scale and array or non-array placement style, we correspondingly use some adaptive strategies to greatly reduce the runtime. We make innovations in the adaptive strategies for constructing single crossover meme and accepting placement candidate. The experimental tests are performed on Peko suite3 and ISPD04 benchmark circuits, the results and comparisons show that these strategies are efficient.