{"title":"伪随机数生成器对遗传算法性能的影响以最小化海上货物交付路线长度","authors":"V. Romanuke, Andriy Romanov, Mykola O. Malaksiano","doi":"10.31217/p.36.2.9","DOIUrl":null,"url":null,"abstract":"We consider a problem of minimizing the maritime cargo delivery route length to reduce the delivery cost. In our model, the cost is equivalent to the sum of tour lengths of feeders used for the delivery to cover the route. Formulated as a multiple traveling salesman problem, we solve it with a genetic algorithm. The algorithm performance is dramatically influenced by the stream of pseudorandom numbers used for randomly generating the starting population and accomplishing random mutations. As the number of ports increases from 10 to 80, the route length variation intensifies from 3.5% to 22.5% on average. However, we increase the route length minimization accuracy by re-running the algorithm to solve the same problem until closely the best solution is obtained. The number of reruns is about 3 to 6 for up to 20 ports. For more than 20 ports the required number of algorithm reruns abruptly increases from 28 reruns for 30 ports to about 51 reruns within the range of 40 to 80 ports.","PeriodicalId":44047,"journal":{"name":"Pomorstvo-Scientific Journal of Maritime Research","volume":" ","pages":""},"PeriodicalIF":0.5000,"publicationDate":"2022-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Pseudorandom number generator influence on the genetic algorithm performance to minimize maritime cargo delivery route length\",\"authors\":\"V. Romanuke, Andriy Romanov, Mykola O. Malaksiano\",\"doi\":\"10.31217/p.36.2.9\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"We consider a problem of minimizing the maritime cargo delivery route length to reduce the delivery cost. In our model, the cost is equivalent to the sum of tour lengths of feeders used for the delivery to cover the route. Formulated as a multiple traveling salesman problem, we solve it with a genetic algorithm. The algorithm performance is dramatically influenced by the stream of pseudorandom numbers used for randomly generating the starting population and accomplishing random mutations. As the number of ports increases from 10 to 80, the route length variation intensifies from 3.5% to 22.5% on average. However, we increase the route length minimization accuracy by re-running the algorithm to solve the same problem until closely the best solution is obtained. The number of reruns is about 3 to 6 for up to 20 ports. For more than 20 ports the required number of algorithm reruns abruptly increases from 28 reruns for 30 ports to about 51 reruns within the range of 40 to 80 ports.\",\"PeriodicalId\":44047,\"journal\":{\"name\":\"Pomorstvo-Scientific Journal of Maritime Research\",\"volume\":\" \",\"pages\":\"\"},\"PeriodicalIF\":0.5000,\"publicationDate\":\"2022-12-23\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Pomorstvo-Scientific Journal of Maritime Research\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.31217/p.36.2.9\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q4\",\"JCRName\":\"TRANSPORTATION\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Pomorstvo-Scientific Journal of Maritime Research","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.31217/p.36.2.9","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"TRANSPORTATION","Score":null,"Total":0}
Pseudorandom number generator influence on the genetic algorithm performance to minimize maritime cargo delivery route length
We consider a problem of minimizing the maritime cargo delivery route length to reduce the delivery cost. In our model, the cost is equivalent to the sum of tour lengths of feeders used for the delivery to cover the route. Formulated as a multiple traveling salesman problem, we solve it with a genetic algorithm. The algorithm performance is dramatically influenced by the stream of pseudorandom numbers used for randomly generating the starting population and accomplishing random mutations. As the number of ports increases from 10 to 80, the route length variation intensifies from 3.5% to 22.5% on average. However, we increase the route length minimization accuracy by re-running the algorithm to solve the same problem until closely the best solution is obtained. The number of reruns is about 3 to 6 for up to 20 ports. For more than 20 ports the required number of algorithm reruns abruptly increases from 28 reruns for 30 ports to about 51 reruns within the range of 40 to 80 ports.