Cyclic robotic cells scheduling using tabu search algorithm

A Robotic cell is a common complex manufacturing system with two main characteristics: (a) material handling operations in the system are performed by a computer-controlled robot; (b) the processing time on each workstation is flexible, which can be an arbitrary value within a given time window. This paper presents a tabu search (TS) algorithm to find an optimal cyclic schedule in a robotic cell with the objective of minimizing the cycle time or equivalently maximizing the throughput. The schedule in such a complex system is featured by the sequence of robot moves and their starting times. We first formulate the problem provided that the robot move sequence is given. Then, we divide the whole solution space of the problem into a series of sub-spaces by the works-in-process (WIPs) levels in the robotic cell. Then we propose a TS algorithm to search the best solution in each subspace synchronously. In the proposed TS algorithm, we design a constructive heuristic to generate initial solutions and use a swapping operation to reach the neighborhood. The computational results conducted on benchmark problems indicate that the proposed algorithm is effective for the cyclic robotic cell scheduling problem.