MDP based optimal control for a colloidal self-assembly system

Markov decision process (MDP) based optimal control for colloidal self-assembly systems is proposed in this paper. The colloidal assembly process is interpreted as a Markov process. Starting with state space discretization, the system's dynamics under different actuator conditions are modeled by actuator-parameterized Markov chain models. Then the MDP based optimal control is applied to achieve design goals. This dynamic programming approach is illustrated through a simulated depletion colloidal assembly system. Simulations show that the proposed MDP based optimal control is able to effectively drive the system to the desired state, i.e. the state corresponding to the desired perfect crystal structure.