Genetic Programs and Co-Evolution: Developing robust general purpose controllers using local mating in two-dimensional populations

A co-evolutionary approach for developing programs for controlling a very simple \robot-like" simulated vehicle is presented. The main goal is to nd programs that can generalize and solve other similar problems. Good results are achieved by coevolving the test cases and the simulated vehicles and using locality in both the reproduction and evaluation phases. The tness of a controller is determined by its performance in competition with its neighbours in the test case population. The tness of a test case is similarly determined through competition with its neighbours in the controller population. The co-evolved controllers are more robust and general than a simple hand-designed algorithm or controllers evolved using a xed training set.