Representation of solutions in genetic VLSI placement algorithms

Abstract

The VLSI placement problem is presented in this article. A mechanism of representation of solutions for further genetic algorithm implementation is described. The proposed encoding algorithm is based on a placement tree and reverse Polish notation. The decoding algorithm is implemented in two stages: twinning of elements in macroblocks and calculation of real coordinates of elements. Experimental results show time-response characteristics of the proposed coding and decoding mechanisms. The time complexity of the encoding algorithm is represented by O(n) whereas the time complexity of the decoding algorithm is represented by O(n log n), where n is the number of elements.