Simulation of using of markov decision processes in restaurant robots

In order to make restaurant services more intelligent and complete basic services, the robot’s forward path needs to be planned. The whole restaurant could be regarded as a probabilistic model and Markov decision process (MDP) can help solve this problem. This paper proposes has a more appropriate approach - Stoical shortest path (SSP), a form of MDP. SSP can update the value of the paths iteratively and compare them to find the path with the lowest cost and could instruct the shortest way for the robot.