We reviewed one important approach to the classical MDVRP optimization – GA. As we saw in the background section, there has been already a lot of research done in this field, but none of them review the field systematically. We tried to fill this gap, by reviewing the most important achievements done in the field. Most frequent methods of crossover, mutation and selection were described, implemented and also tested later on. We also review some of the most used non typical operators, like elitism and repair operators. Based on the literature review, some common and most useful genetic operators were implemented and run through the experiment in order to compare them to each other and find the most useful ones. As we saw from the results of the experiment, some approached were not particularity successful in providing competitive solutions for this kind of problem. Others, like ordered crossover, exchange mutation and tournament or linear ranking selection were found to be quite good and should be used in any further study done with the similar type of the problem in mind.
Some of the leading GA approaches to solving MDVRP were tested on standard benchmark problems and compared with other existing MDVRP optimization methods. The obtained results show, that the GAs are on par with other approaches and are very appropriate for this type of problem solving. As has also been confirmed in [18] and [19], the main advantage of the GA is the linear scaling with the growing problem size and thus is preferred for solving large NP problems over exact algorithms or some other heuristic methods.
We reviewed one important approach to the classical MDVRP optimization – GA. As we saw in the background section, there has been already a lot of research done in this field, but none of them review the field systematically. We tried to fill this gap, by reviewing the most important achievements done in the field. Most frequent methods of crossover, mutation and selection were described, implemented and also tested later on. We also review some of the most used non typical operators, like elitism and repair operators. Based on the literature review, some common and most useful genetic operators were implemented and run through the experiment in order to compare them to each other and find the most useful ones. As we saw from the results of the experiment, some approached were not particularity successful in providing competitive solutions for this kind of problem. Others, like ordered crossover, exchange mutation and tournament or linear ranking selection were found to be quite good and should be used in any further study done with the similar type of the problem in mind.บางวิธี GA นำไปแก้ MDVRP ถูกทดสอบบนปัญหามาตรฐานมาตรฐาน และเปรียบเทียบกับวิธีการเพิ่มประสิทธิภาพ MDVRP อื่น ๆ ที่มีอยู่ ผลได้รับแสดง ก๊าซจะไล่เลี่ยกับวิธีอื่น ๆ และเป็นที่เหมาะสมสำหรับชนิดของการแก้ปัญหานี้ ขณะที่ยังได้รับการยืนยันใน [18] [19], ประโยชน์หลักของ GA เป็นเส้นขนาดกับขนาดปัญหาเจริญเติบโต และจึง เป็นที่ต้องการแก้ไขปัญหา NP ใหญ่อัลกอริทึมที่แน่นอนหรือบางวิธีอื่น ๆ แล้ว
การแปล กรุณารอสักครู่..