Ceļojošā tirgotāja problēmas risinājumu metožu salīdzinājums
Loading...
Date
Authors
Advisor
Journal Title
Journal ISSN
Volume Title
Publisher
Latvijas Universitāte
Language
lav
Abstract
Darbs ir veltīts Ceļojošā tirgotāja problēmas padziļinātai izpētei, tās risināšanas īpašī-bām un sarežģījumiem, kas ir saistīti ar to. Darbā tiek definēta Ceļojošā tirgotajā problēmas nostādne, tās veidi un pieeja risināšanai. Darbā tiek aplūkoti vairāki algoritmi Ceļojošā tirgotāja problēmas risināšanai, to teo-rētiskais pamatojums, sagaidāmā efektivitāte, to izpildes laiks, kā arī sarežģījumi to pielie-tošanā. Darbā tiek praktiski pielietoti vairāki Ceļojošā tirgotāja problēmas atrisināšanas algoritmi publiski pieejamiem piemēriem, kuriem arī ir zināms optimālais risinājums. Gal-venais darba uzdevums ir noskaidrot, kuri no publiski pieejamiem algoritmiem ir visefektī-vākie un kādos gadījumos tie dos vislabākos rezultātus, kā arī tika apskatīti to uzlabošanas mehānismi. Risināšanas metodes tika sadalītas tiešajās un tuvinātās. Algoritmu palaišanas rezultējošo ceļu garumi un laiki tiek apkopoti failos, kas pēc tam tiek izpētīti, analizēti un attēloti grafikos programmā R. Rezultātiem tika piemērotas vairākas statiskās analīzes, kas ļauj precīzāk noteikt algoritmu izpildes efektivitāti. Rezultā-tā tiek apkopota algoritmu efektivitāte un analizēta to pielietošanas nepieciešamība un ie-spējamība. Atsevišķi tika apskatīts Concorde algoritms, kas tiek uzskatīts par visefektīvāko metodi.
The paper is devoted to deeper analysis of the comparison of the methods of solving Travelling Salesman Problem, which mainly includes the approaches of solving the instanc-es of the Travelling Sales Problem, the properties and the possible difficulties which may be faced. There are several Traveling Sales Problem solving algorithms analysed in the paper, with the theoretical explanation, expected time complexity and the difficulties when using them. Some of the algorithms analysed are executed on the publicly available instances, which have the Optimal solution already known. The main goal of the paper is to check the efficiency of different algorithms and find out in which cases they give the best results and the possibility to improve them. The algorithms were split into the exact and approximate approaches. The results of the execution of different methods are written into the files which are then analysed and graphically displayed using R program. There are multiple statistical characteristics calculated and the analysis done to show the efficiency of different algo-rithms. All the results are summarized to determine the efficiency of the algorithms applied. Separately, there is Concorde algorithm analysed, which is considered as the most effective solver of the Travelling Salesman Problem.
The paper is devoted to deeper analysis of the comparison of the methods of solving Travelling Salesman Problem, which mainly includes the approaches of solving the instanc-es of the Travelling Sales Problem, the properties and the possible difficulties which may be faced. There are several Traveling Sales Problem solving algorithms analysed in the paper, with the theoretical explanation, expected time complexity and the difficulties when using them. Some of the algorithms analysed are executed on the publicly available instances, which have the Optimal solution already known. The main goal of the paper is to check the efficiency of different algorithms and find out in which cases they give the best results and the possibility to improve them. The algorithms were split into the exact and approximate approaches. The results of the execution of different methods are written into the files which are then analysed and graphically displayed using R program. There are multiple statistical characteristics calculated and the analysis done to show the efficiency of different algo-rithms. All the results are summarized to determine the efficiency of the algorithms applied. Separately, there is Concorde algorithm analysed, which is considered as the most effective solver of the Travelling Salesman Problem.