Evolving predator and prey behaviours with co-evolution using genetic programming and decision trees

Abstract

Developing artificial behaviours to control artificial creatures or vehicles is a task that can be solved by means of Evolutionary Algorithms. The Predator and Prey is a problem where it is possible to evolve behaviours for both predator and prey, using artificial co-evolution: the predator must capture the prey and the prey must evade the predator. Both predator and prey have also different characteristics, the predator is faster and more agile and the prey is slower. This paper presents an alternative, using Genetic Programming with Decision Trees for evolving both Predator and Prey behaviours. The results obtained shows the feasibility of the approach.

Automatic Generation of 2-AntWars Players with Genetic Programming

“…Both players control an ant. The ant of player 1 is located at position (2, 5) and the ant of player 2 at (8,5). The aim of the game is to collect more of the 15 available pieces of food than the opponent.…”

“…Hunting would not be possible on a toroidal field because the hunted ant could flee indefinitely. Each of the two players has control over two ants, which start at positions (0, 5) and (0, 7) for player 1 and at (19,5) and (19,7) for player 2. Every ant has the same capabilities as their AntWars brethren, i.e.…”

“…In [4] space combat strategies were created. Other forms of predator-prey interaction were analyzed in [5] and [6]. Genetic programming has also been used to develop soccer [7] and chess end game players [8].…”

Evolution for modeling

“…In [6], the feasibility of using GP to create agent behavior is demonstrated in a Predator and Pray scenario. Agents derive their decisionmaking process from a decision tree model, that is built throughout the execution.…”

Evolving combat algorithms to control space ships in a 2D space simulation game with co-evolution using genetic programming and decision trees

“…We also evolved behaviours using other attack and flee trees [4], and the results achieved the main goal while employing tactics that took advantage of the attack and flee trees strong points. Even the worst attack and flee trees could be used to achieve the main goal.…”