A two-stage no-wait job shop scheduling problem by using a neuro-evolutionary variable neighborhood search

چکیده مقاله

A two stage order scheduling problem with an additional no-wait condition is studied in this article. All orders must be processed from start to completion without any waiting between operations and any interruption within operations. Examples of such a problem occur in steel production, and pharmaceutical, chemical, and construction industries. A new optimization method which is based on a combination of an enhanced Variable Neighborhood Search (VNS) and an Artificial Neural Network (ANN) is devised in this paper. The aim is to simultaneously employ the key characteristics of two approaches to achieve superior solutions in solving the addressed optimization problem. In proposed algorithm, the VNS performs a global search whereas the ANN plays the role of intensive initial solution. The individual obtained by feed forward back-propagation ANN is improved within enhanced VNS iterations. The two techniques complement each other when ANN feeds its best solution to the VNS algorithm. Furthermore to establish the best sequence of order pairs, a Team Process Algorithm is also embedded into the VNS to enhance the search diversification as well as its intensification. The performance of proposed Neuro-Evolutionary algorithm is investigated via computational experiments, and the results prove the appropriate performance of suggested VNS-ANN algorithm.