Hopfield신경망에 의한 주문생산 일정계획

Job shop scheduling for hopfield neural networks

Job shop scheduling belongs to a called Constraint optimization problems. Scheduling involves looking for a combination that minimizes a cost function subject to the given constraints of the problem. For the constraing optimization problems dealing with discrete of continuous parameters, it is desirable to find the optimal value such that energy function, Lyapunov function can be minimized. In job shop scheduling, these are n jobs with m machines, it is desirable to find the best job sequence in scheduling the n jobs on the m machines so that the earliest completion time can be achieved. Since job shop scheduling have been shown to be NP-complete problems, this means that the computation requirement does increase in a polynomial fashion. For NP-complete problems, the optimal shlution in not always obvious and cannot be guaranteed unless the entire solution base is known. In this paper, the simulated annealing approach is applied for the minimization of energy function. Experiment results, it can be speeded up the convergence rate of the energy function.