Job-shop Scheduling And Its Applications Based On Constraint Satisfied NN

Production scheduling is very popular in social lives, with an intensive background of practical applications in such areas as industry, agriculture, national defence, science research, transaction and services. Manufacturing is so important and basic for a nation's economy that a country can't be truly flourishing and strong without developed manufacturing. The production scheduling methods are one of the most important manufacturing techniques. And of all scheduling problems, the job-shop scheduling problem (JSP) is most common and complicate, which is usually a typical NP-complete problem. So it is of great importance, both theoretically and practically, to develop job-shop scheduling methods. There are few fruits in the fields of JSP. We can find few effective methods in very specific condition. The methods for JSP can be divided into 3 classes, which are Accurate Optimal Approach, Approximally Heuristic Approach and Local Search Based Approach.With the rapid development, Artificial Intelligence has also been widely used to solve optimization and scheduling problems. Artificial neural network is a new active member of all intelligent optimization methods. Based on previous constraint satisfaction neural networks, this thesis applies a new constraint satisfaction neural network (CSNN) for job-shop scheduling problems.Based on the general model of nerve cell, the CSNN declares 3 types of nerve cells as following: (1) ST (Starting Time) cells, (2) SC (Sequence Constraint) cells and (3) RC (Resource Constraint) cells. STs regard the starting time of an operation. SCs and RCs regard the sequence and the resource constraint between operations, while RCs regard the resource constraint.There are 3 types of executing mechanism for the CSNN, which are...