Large-scale multi-agent system has been a popular research field in distributed artificial intelligence. A major challenge exists in team overall performance optimization because there are a large number of uncertainties in agents' complex processes of interactions. In our previous work, we have found that a small change of a parameter may influence the performance significantly. On the other hand, existing researches of large-scale multi-agent coordination algorithm only proved the evidences by simply comparing with the non-intelligent coordination method, where a general simulation platform is required that can abstract the details of environment but is able to simulate the complex coordination activities for each agent. In this paper, we have designed such a general platform for large-scale multi-agent team coordination. This platform is able to simulate a variety of tasks that are imperative to agents' team coordination. In addition, by varying the setting of giving team coordination parameters as well as loading different coordination strategies, it is capable of simulating and evaluating system performance according to different properties and provides the clue on building optimization algorithms in future.
