An adaptive feedback neural network approach to job-shop scheduling problem

Job-shop scheduling problem is a typical representative of NP-complete problems and it is also a popular topic for the researchers during the recent decades. Lots of artificial intelligence techniques were used to solve this kind of problems, such as: genetic algorithm, tabu searching method, simulated annealing and neural network. Based on the previous research of Zhou and Willems, this paper proposes a neuro-dynamic model with two heuristics to solve job-shop scheduling problems. The stability of this neural network is proven by using Lyapunov stability theorem. Both small-size and big-size problems are used to test this neural network. Simulation results of some tested samples are given. And the performance of this neural network is compared with several other neural works under experimental conditions.