Distributed autonomous agents, navigation and cooperation with minimum intelligence in a dynamic warehouse application

This paper describes problems of autonomous agents in a multi-agent environment that have only local sensory data. This limitation in sensory data prohibits the agents to anticipate other moving agents in motion planning. Increasing, the performance of the strategy for motion planning, together with the population density of moving cells, results in decreasing the performance of the navigation process. This decrease is caused possibly due to counteracting influences of the strategy and population density. Experiments on a cellular warehouse application show that some of the characteristics of counteracting influence are emerging, but the relatively great number of collisions has a far greater influence on the performance of the warehouse. This masks the effect of the counteracting influence.