Application And Research Of Scheduling Production With Intelligent Approach

Scheduling approach is a kind of portfolio optimizations which belong to NP category. It is difficult to find the best result to the scheduling problem. We should try to find an approximate and useful result in a reasonable and limited time instead of finding the best one, which is the best algorithm with most engineering significance. For recent years, various intelligent algorithms have been introduced to the field of scheduling problems, such as Genetic Algorithm, Simulated Annealing Algorithm, Taboo Search Algorithm and so on.Genetic Algorithm, combining simulating genetics and theory of nature-selecting construction, is a kind of Search Algorithms. Because it rarely dependent on the optimizing problems and is characteristic of non-linearity, robustness and implicit parallelism, it is widely applied to relative fields. Based on the application of Genetic Algorithm to producing schedule, the paper elaborates three points as following:1. Job-shop scheduling problem (JSSP) based on Genetic AlgorithmJSSP is a typical NP- problem. Aimed at the particularities of JSSP, the paper studies the code/decode approach to Genetic Algorithm, the design of genetic operators and the change of objective functions and fitness proportion. The author also designs a kind Genetic Algorithm for JSSP and simulates with famous Fisher and Thompson's Bench Mark, furthermore, the author designs an improved Genetic Algorithm for JSSP.2. Flow-shop scheduling problem (FSSP) based on Genetic AlgorithmFSSP is a kind of complex and typical assembly line scheduling problems, which can be solved by Genetic Algorithm successfully. The problem of assembly line balance, computed by Ranked Position Weighted Method, has been considered a plan problem up to now; however, the author thinks that it should be a dynamic production scheduling problem. Therefore, a hybrid genetic algorithm is designed and the balance problem is solved effectively.3. Multi-processor scheduling problem and Genetic AlgorithmThe essence of solving Multi-processor scheduling problems lies in the reasonable distribution of all work pieces and the sequence of processing. The paper introduces the parallel machine's early/tardy schedule, then designs the Genetic Algorithm for parallel machine schedule with process constraint and stimulates through examples.In the field of scheduling approach, although Genetic Algorithm is applied widely, its shortcomings are also obvious. What the main field studied in for a long time is to accelerate algorithm's convergence rates and precision. Because various improved algorithms (Adaptive GA, Simulated Annealing GA, Parallel GA and Orthogonal GA) have their advantages at one aspect respectively, how to accelerate evolutive speed and improve the ability to fight premature convergence are the problems which people expect to solve.