| Summary: | 碩士 === 龍華科技大學 === 商學與管理研究所碩士班 === 93 === Owing to the rapid changing of competitive environment and consumer’s need, enterprise must improve product quality and after sale service at any time. In the meantime, enterprise must construct quick response required alternation model, and change existent operational method and order so as to quickly but effectively response this situation. Most of the researches of scheduling problem only aim directly at single objective which leads optimal solution or best scheduling order. In fact, for enterprise, the schedule of the production flow must consider many factors. So it is a must to research multiple objective integrated production scheduling problem. This research takes flow shop production scheduling as its range, making the most of genetic algorithm to quickly seek best approximate solution. The expectant aim of this research includes： to construct a scheduling model, enterprise can solve scheduling problem under different operation objectives and condition and exert construct genetic algorithm model；to quickly find out optimal production scheduling order, and to take advantage of scheduling rules combination, and with these aims enterprise can fully go with environmental alteration.
In the part of data verify, the result of 300 generations algorithm can concluded as follows： in tardiness time, as for the result of single objective scheduling planning, the improved effect of initial solution 996 decrease to GA optimal solution 235 is 75.67% that proves GA is very effective theory for search. As for the multiple objective scheduling planning, in tardiness time and job waiting time, the improved effect of initial solution 829 decrease to GA optimal solution 240 is 71.05% in the tardiness time; the improved effect of initial solution 2589 decrease to GA optimal solution 915 is 64.66% under the same condition in the job waiting time. In the special problem scheduling planning, GA optimal solution proceed comparison between considering specific equipment situation and no considering specific equipment situation, leading to makespan for 1096 and machine idle time for 409 in the single objective, in the other hand machine idle time falls from 103 to 0 in the specific equipment, makespan for 1110 of specific equipment, only little increased by 1.26%. Verify data mentioned above illustrates that the scheduling model presented by this research can in effect settle many kinds of scheduling problems in the different situation, and with excellent algorithm effect.
|
|---|
