The Player Coalition Strategy Research Of MMOG

In the MMOG(Massively Multiplayer Online Games), the collaboration and interaction among the players make MMOG more competitive and interesting. But in the current MMOG, the player coalition strategy applied is relatively easy, players form the coalition in a simple way to complete the task that can't be completed by the single player, the player can not be choosed according to the mission requirements. Therefore, it is great significant to study the players coalition strategy which is suitable for the MMOG.This paper presents a method for player coalition strategy which is suitable for MMOG, through research the related technologies of Agent coalition strategies and methods. This paper completd the following two areas:The paper proposes the dynamic coalition strategy based on the Contract Net Negotiation Mechanism to solve the problems of the player coalition strategy in MMOG at present. This method brings to the player comprehensive assessment mechanism, which is capable of the player ablity value, the historical experience, and the credit values, the coordinator computes the comprehensive assessment value and chooses the coalition candidate players by the value in order to reduce the traffic during the coalition, and meanwhile the coordinator negotiates with the candidate members using the improved dynamic contract net mechanism. The coordinator forms the players coalition which is suitable for the task by the improved ant colony algorithm, when received sufficient bidding of the player. By improving the heuristic function of the ant colony algorithm, and set different weights in the players ability value and the SEC value of the heuristic function, the coordinator give priority to choose the player who previously worked with and suitable for the task to form the coalition.For the enemy events need link with the battlefield environment in the games, to provide the auxiliary decision for the players attacking-target decision using Bayesian network. In coalition combat, how to select the attacting-target enemy, making the coalition as a whole more effective has become an important issue in coalition operations. It is very difficult to select the target enemy from the complex information when there are many enemies on the battlefield, so it is nessesary to provide the auxiliary decision for the player selecting the target. The Bayesian network has many advantages in dealing with the probability issue, Therefore, this paper studies the principle of Bayesian network with the characteristics of the online games, establishes Bayesian model for the player attacking-target using Bayesian network, and then calculate the strengths and weaknesses of the probability of the player attacking-target through the Bayesian probability formula, to provide the auxiliary decision-making for the player.This paper carried out the experiment based on robocode simulation tank engine, the experiment results show that this algorithm can effectively reduce the search time, the amount of computation and the traffic, so it can improve the capability of cooperation and campaign over all the team.