Using Artificial Intelligence Algorithms for the Museum Routing Problem with Immediate Visiting and Capacity Constraints

• Main Authors: HUANG, YU-LUN, 黃郁倫
• Other Authors: HSIEH, YI-CHIH
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/tu34c7

碩士 === 國立虎尾科技大學 === 工業管理系工業工程與管理碩士班 === 107 === With the development of history, in order to avoid the destruction of historical relics and preserve cultural heritage, through the museum, some cultures can be preserved and tourism benefits can be created and entertainment, cultural landscapes and education and learning opportunities can be provided through collection, preservation and restoration. Visiting museums are often one of the must visit spots for people to travel and one of the most common off-campus studies. They can learn more about historical relics and historical artifacts. In order to avoid the poor quality of visits caused by the congestion of a large number of tour groups during the peak season, planning the museum visiting path is relatively important. In this study, the study seeks to planning the museum visiting path for all tour groups from the same organization such that the total completion time of all tours is minimized. This study explores a new museum visitor routing problem in which all exhibition rooms are divided into two categories, namely the must-visit exhibition room and the select-visit exhibition room. The former is the exhibition room that all tour groups must visit, and the latter is the exhibition room that tour groups can visit or not visit. In this study, we assume that (a) all exhibition rooms are immediate visiting that means that all rooms can be visited at any time, and (b) the capacities of all exhibition rooms are limited. The objective of this new museum visiting problem is to schedule the visit paths for all tour groups from the same organization such that the makespan of all tour groups is optimized. The problem studied in this thesis is an extension the Open Shop Scheduling Problem (OSSP) and it is an NP-hard problem. This study takes Taiwan Literature Museum (National History Museum) as an example and uses Genetic Algorithm (GA), Immune Algorithm (IA) and Particle Swarm Optimization (PSO) to solve. Additionally, we propose a new encoding scheme to convert any random integer sequence into a feasible solution of the problem. Numerical results of this study show that Immune Algorithm to are superior to the other two algorithms in solution quality.