Strategy selection in games using co-evolution between Artificial Immune Systems

In this paper, we create a simple artificial computer game in order to illustrate a means of making software players in computer games more intelligent. Despite being known as AIs, the current generation of computer players are anything but intelligent. We suggest an algorithm motivated by concepts from Artificial Immune Systems (AIS) in which an attack from one opponent is met with a response from the other which is refined in time to create an optimal response to that attack.. We refine the AIS algorithm in that we model the response with a probability vector rather than a population of antibodies. Some typical results are shown on the simple computer game.