A novel agent bidding based optimization approach in manufacturing planning and scheduling

The problem of process planning and scheduling in manufacturing systems can be simply concluded as an optimization problem in terms of minimizing processing cost and time with the constraints of different manufacturing resources capabilities and capacities and fixed operating sequence of job features. Some previous literatures involve centralized optimization approaches which are able to find a close to optimal solution, but the computational complexity is extremely high and the model is not flexible. Some others adopted dispatching rules, which can achieve good solutions, but not optimal ones. This paper presents a novel agent bidding based mechanism to solve the problem. A weight function is introduced to facilitate the search for optimal allocation, where the weight factors control the preference of production cost and time. Meanwhile, in each bidding round, there is an exploration process which gives small possibility to the agents with non-optimal bids to be selected as the winner. After iterations for adjusting parameters and exploration, the near optimal solution with minimized cost and satisfied lead time can be found.