Evolutionary Multiobjective Optimization on a Chip

The majority of research in evolvable hardware is focused on evolving logic for deployment on reconfigurable hardware. There are far fewer reports concerned with the implementation of evolutionary algorithms (EAs) in hardware. The focus of our research is directed toward using reconfigurable hardware as a means to speed up evolutionary search, and in particular evolutionary multiobjective optimization (EMO). Evolutionary multiobjective optimization utilizes an evolutionary search to find solutions to difficult multiobjective optimization problems. We present an implementation of an EMO algorithm in reconfigurable hardware, and discuss how it may be utilized in practical deployment situations