期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2017
卷号：95
期号：2
出版社：Journal of Theoretical and Applied
摘要：Recently, there has been a growing attention to employ evolutionary algorithms (EAs) in addressing dynamic optimisation problems (DOPs) due to its significance in real world applications. The most notable challenge when solving DOPs is that the objective should not only attempt to seek the global optimum by an efficient way, but be able to keep track the optimal solution during the environmental changes. Thus, several mechanisms have been developed for EAs in order to improve the search performance of the algorithm in accommodating the dynamic changes such as by increasing the diversity of the population. Among these strategies, the multi-population mechanism has been found beneficial for EAs for DOPs. Dynamic travelling salesman problems (DTSPs) are categorised under DOPs. In the Travelling Salesman Problem (TSP), a salesman wants to distribute items sold in different cities starting from his home city and returning after he visited all the cities to his starting city again by optimising his time and tour efficiently. However, in the DTSPs, it is more challenging to consider the traffic delays that may affect the route of the salesman and change the time planned beforehand. Therefore, the salesman will optimise his time again and find a new alternative route to avoid long traffic delays. The presented work aims to build upon the state of the art research methodologies for the DTSPs with traffic factors, where in order to cope with the dynamic behaviour, a multi-population approach is applied to harmony search algorithm that mimics the musical process of trying to find a state of harmony. Moreover, a multiple pitch adjustment rate (PAR) strategy is proposed since PAR assumed to be the moving rate from one city to the nearest city in the TSP. The performance of the proposed multi-population HS algorithm is verified on two variations of DTSPs with traffic factors, i.e., random and cyclic traffic delays. Based on different DTSP test cases, the experimental results show that the proposed approach is able to obtain competitive results when compared to the best-known results in the scientific literature.
