EMO'07 Proceedings of the 4th international conference on Evolutionary multi-criterion optimization
| Hi-index | 0.00 |
This paper describes a multi-objective Evolutionary Algorithm (EA) system for the synthesis of efficient non-linearVLSI circuit modules. The EA takes the specification fora non-linear block, and converts it into a technology independentnetlist, specified in the Verilog hardware descriptionlanguage. The hardware designs are based upon highlevelcomponents such as adders and multipliers. The circuitdesigns that are produced are near-optimal with respectto silicon area and longest-path delay.The performance of the EA is enhanced through the useof local searches. These searches are embedded withinthe genetic operators, and enable the rapid evaluation oflarge numbers of designs. The use of searches increases thepower of the EA system, without forfeiting the benefits ofusing a population of solutions.The system is demonstrated with several test problems.Results are presented for the discovery of correct designs,and also regarding the quality of the evolved designs. Themost complex designs have areas as large as 200,000_m2in a 0.18_m technology.