Output details
15 - General Engineering
Brunel University London
Binary string fitness characterization and comparative partner selection in genetic programming
The survival of the fittest characterizes evolutionary computational methods, requiring fitness measures for individuals. This paper invents novel strategies for evaluating individual’s relative strengths and weaknesses, and representing them in a fundamentally new binary string fitness characterization (BSFC). A new rigorous paradigm is created by utilizing the BSFC in proposing a pair-wise mating strategy, Comparative Partner Selection, in evolving a population that promotes effective solutions by reducing population-wide weaknesses. Published in a high impact factor journal, this represents a significantly promising development that subsequently led to successes in breast cancer detection, communications (IEEE TWC 2012), and condition monitoring applications.