Chaiya Chomchalao

This study aims to select the ideal mixture of small and medium-sized destinations and attractions in Thailand’s Ubon Ratchathani Province in order to find potential wellness destinations and attractions. In the study region, 46 attractions and destinations were developed as the service sectors for wellness tourism using the designed wellness framework and the quality level of the attractions and destinations available on social media. Distinct types of tourists, each with a different age and gender, comprise a single wellness tourist group. Due to them, even with identical attractions and sites, every traveler has a different preference. A difficult task for travel agencies is putting together combinations of attractions and places for each tourist group. In this paper, the mathematical formulation of the suggested problem is described, and the optimal solution is achieved using Lingo v.16. Unfortunately, the large size of test instances cannot be solved with Lingo v16. However, the large-scale problem, particularly the case study in the target area, has been solved using a metaheuristic method called AMIS. According to the computation in the final experiment, AMIS can raise the solution quality across all test instances by an average of 3.83 to 8.17 percent. Therefore, it can be concluded that AMIS outperformed all other strategies in discovering the ideal solution. AMIS, GA and DE may lead visitors to attractions that generate 29.76%, 29.58% and 32.20%, respectively, more revenue than they do now while keeping the same degree of preference when the number of visitors doubles. The attractions’ and destinations’ utilization has increased by 175.2 percent over the current situation. This suggests that small and medium-sized enterprises have a significantly higher chance of flourishing in the market.

This paper presents an algorithm to solve the multilevel location–allocation problem when sabotage risk is considered (MLLAP-SB). Sabotage risk is the risk that a deliberate act of sabotage will happen in a living area or during the transportation of a vehicle. This can change the way decisions are made about the transportation problem when it is considered. The mathematical model of the MLLAP-SB is first presented and solved to optimality by using Lingo v. 11 optimization software, but it can solve only small numbers of test instances. Second, two heuristics are presented to solve large numbers of test instances that Lingo cannot solve to optimality within a reasonable time. The original differential evolution (DE) algorithm and the extended version of DE—the modified differential evolution (MDE) algorithm—are presented to solve the MLLAP-SB. From the computational result, when solving small numbers of test instances in which Lingo is able to find the optimality, DE and MDE are able to find a 100% optimal solution while requiring much lower computational time. Lingo uses an average 96,156.67 s to solve the problem, while DE and MDE use only 104 and 90 s, respectively. Solving large numbers of test instances where Lingo cannot solve the problem, MDE outperformed DE, as it found a 100% better solution than DE. MDE has an average 0.404% lower cost than DE when using a computational time of 90 min. The difference in cost between MDE and DE changes from 0.08% when using 10 min to 0.54% when using 100 min computational time. The computational result also explicitly shows that when sabotage risk is integrated into the method of solving the problem, it can reduce the average total cost from 32,772,361 baht to 30,652,360 baht, corresponding to a 9.61% reduction.