Optimization model for designing personalized tourism packages. (January 2023)

• Record Type: Journal Article
• Authors: Piya, Sujan; Triki, Chefi; Al Maimani, Abdulwahab; Mokhtarzadeh, Mahdi
• Is Part Of: Computers & industrial engineering. Volume 175(2023)
• Publication Date: 2023-01
• Subjects: Tourism industry -- Tourist preferences -- Multi-objective optimization -- Meta-heuristic algorithms; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2022.108839
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: Personalized Tourism Package is multi-objective in nature. Optimization models for Optimum Personalized Tourist Package is developed. NASG-II and NASG-III outperform MOGWO and MOICA heuristics. NSGA-II is better in terms of Pareto optimal solutions and solution time. NSGA-III is better at exploring solution space. Abstract: The tourism supply chain aims at satisfying the needs of the tourists based on their preferences. However, the preference of each tourist may be different. Some tourists prefer to optimize a single criterion, while others prefer to optimize conflicting multiple-criteria. The tourism service provider can hardly offer the tourists with the itinerary according to their precise preferences. This paper proposes a multi-objective optimization framework based on which tourists can generate itineraries according to their preferences. A mathematical model is presented, which is multi-objective and NP-hard. Consequently, four meta -heuristic algorithms, namely none-dominated sorting genetic algorithm versions II (NSGA-II) and III (NSGA-III), multi objective grey wolf optimization, and multi objective imperialist competitive algorithm are developed. The proposed method helps the tourists to compare different combinations of activities and select the one that best suits their preferences. The model is tested on a small-scale real case pertaining to the Sultanat of Oman. Thereafter, the performances of the proposed algorithms were evaluated on large scale problems. The result shows that NSGAs outperformed other algorithms. NSGA-II outperformed its NSGA-III counterpart in small instances. Surprisingly, as the size of the problem increases, the efficiency of NSGA-II decreases while that of NSGA-III increases. In large instances, NSGA-III outperformed NSGA-II.