We have developed a technique for allocation and binding for data path synthesis using a Genetic Algorithm approach. The use of the genetic algorithm is relatively new in EDA tools. The proposed genetic algorithm uses a non-conventional crossover mechanism, relying on a novel force directed data path binding completion algorithm. The developed system has a number of features such as acceptance of some design parameters from the user, use of a bus based topology, use of multiport memories and provision for multicycling and pipelining, among other features. The results obtained on the standard examples have been very promising.

Keywords: Data Path Synthesis, Binding, Multiport Memory, Genetic Algorithm, VLSI.