Abstract
Multi-agent systems (MAS), consist of a number of autonomous agents, which interact with one-another. To make such interactions successful, they will require the ability to cooperate, coordinate, and negotiate with each other. From a theoretical point of view such systems require a hybrid approach involving game theory, artificial intelligence, and distributed programming. On the other hand, biology offers a number of inspirations showing how these interactions are effectively realized in real world situations. Swarm organizations, like ant colonies or bird flocks, provide a spectrum of metaphors offering interesting models of collective problem solving. Immune system, involving complex relationships among antigens and antibodies, is another example of a multi-agent and swarm system. In this chapter an application of so-called clonal selection algorithm, inspired by the real mechanism of immune response, is proposed to solve the problem of learning strategies in the pursuit-evasion problem.
TopBackground
In order to concentrate on the immune and game theoretic techniques, the rest of the chapter will focus on one class of the multi-agent encounters, i.e. pursuit-evasion problems. They are among the most widespread, challenging, and important multi-agent scenarios and represent some of the most significant potential applications for robots and other artificial autonomous agents. In a typical contest of this sort, one or more pursuers chase one or more evaders around until all preys are captured. Models in which pursuit-evasion problems are examined differ in: environment, number of players, agents’ limitations, definition of capture, optimality criterion, space structure etc. (Isler, Kannan, & Khanna, 2004). Various aspects of pursuit-evasion as well as extensive bibliography on this subject can be found in (Sheppard, 1996).
Key Terms in this Chapter
Stochastic Game: A game, where at any point in time the game is in some state. The game transitions to a new state depend on a stochastic function of the previous state and the interactions among the agents
Directed Mutation: A process that is aimed at creating cells with particular features rather then a random set of cells. First the best cell (in terms of interaction strength with the present antigen) is found and afterwards cells with similar features are created.
Repeated Game: A game consisting of a series of interactions among two or more players. After each interaction players may receive some payoff. Unlike a game played once, a repeated game allows for a strategy to be contingent on past moves
Multi-Agent System: A system consisting of many agents, who can perform tasks individually or co-operate in order to achieve a system goal. A very important factor in such systems are agents’ interaction.Agents’ limitations - anything that prevent agents from acting optimally, e.g. limited perception, limited speed
On-Line Learning: A process, in which a system learns and acts simultaneously
Pursuit-Evasion Game: A game in which, predators, or pursuers, chase preys (evaders) around until the preys are captured. Solution constitutes chasing agents’ optimal strategy, which guarantees execution of their task
Agent: An entity, that perceives its environment and acts upon it in order to realize a given set of goals or tasks
Adaptability: The ability to cope with internal or external changes or to adjust itself to dynamic environments or unexpected events.