Cyclic genetic algorithm with conditional branching in a predator-prey scenario

In its traditional form, the cyclic genetic algorithm (CGA) was found to be a successful method for evolving single loop control programs for legged robots. Its major limitation was the inability to allow for conditional branching, which is required for the integration of sensor inputs in the controller. In recent work, we extended the capabilities of CGAs to evolve multi-loop programs with conditional branching. The design proved successful for the evolution of a controller that allowed a robot to efficiently search for a static target in a square area. In this paper we increase the complexity of the experiment and demonstrate the capability of CGAs with conditional branching to generate a controller the predator in a predator-prey scenario.