Coalition Formation for Multi-agent Pursuit Based on Neural Network

Multi-agent Systems (MAS) often needs to pursue a common goal and in order to achieve that they need to make an effective formation. An approach for coalition formation of multi-agent pursuit based on neural network (NN) and agent group role membership function (AGRMF) model is proposed and thus making a novel algorithm called ARGMF-NN. This new algorithm consists of two parts i.e. feature extraction and group generation. First, the layers of feature extraction can abstract the AGRMF feature for all of the groups. Secondly, those features will be fed to the group generation part based on self-organizing map (SOM) layer which is used to group the pursuers with similar features. Besides, we also come up with the group attractiveness function(GAF) which is used to evaluate the quality of coalitions and the pursuers contribution. It helps in adjusting the main ability indicators of AGRMF and other weights of the whole neural network. The simulation experiment showed that this proposal can improve the effectiveness of coalition formation for multi-agent pursuit and also the ability to adopt pursuit-evasion problem with the scale of growing pursuer team.