Thai Journal of Operations Research : TJOR

A Comparison of the Efficiency of Data Mining Techniques to Create a Model for Forecasting the Use of Cosmetic Products Containing Aloe Vera for Small and Medium-Sized Community Enterprises in Prachuap Khiri Khan Province

Siriruang Phatchuay, Naphattanon Thongpan, Rujira Jullapak — Wed, 20 Dec 2023 00:00:00 +0700

The objectives of this study were to compare the efficiencies of the algorithms and feature selection methods used in data mining techniques in order to create the demand forecasting models for Aloe Vera cosmetic products. The data used in this study consisted of a total of 1,034 data sets. Feature selection was performed in order to select only key features by using CfsSubsetEval and InfoGainAttributeEval in combination with data classification and decision tree techniques. In this study, four techniques were used, including NaïveBayes, MultilayerPerceptron (MLP), J48 and REPTree. The results showed that the use of MLP neural network data mining technique in combination with InfoGainAttributeEval, a feature selection method, showed highest efficiency with an accuracy of 93.54%. The comparison of efficiencies showed that MLP technique can be used in forecasting and as a guide for niche marketing in order to provide advice on purchasing Aloe Vera cosmetic products to interested consumers for better access.

Application of Particle Swarm Optimization to solve Helicopter Arc Routing Problems for High Voltage Transmission Line Inspection

Sippanat Chanthong, Komgrit Leksakul — Wed, 20 Dec 2023 00:00:00 +0700

Electric power is very essential for human daily life. Each year, the demand for electricity is steadily increasing until it affects the production of electricity with limited resources and production potential. Therefore, responsible organizations must balance electricity generation and ensure the safety of electricity consumers. One thing such organizations can do to create safety is to inspect high-voltage transmission lines to ensure they are available and ready to use at all times. By inspecting high-voltage transmission lines, it is necessary to obtain a work plan and minimize operating costs. The capacitated arc routing problem was used as a model for constructing a mathematical model for high-voltage transmission lines, and particle swarm optimization was applied in this research. The mathematical model created can only be used to find answers for small samples and it takes a reasonable amount of time to process the program. However, for large-scale cases and real-world scenarios, it cannot provide answers within a reasonable amount of time. Therefore, in the next step of the research, the cluster-particle metaheuristic method was employed to find additional answers.

The multi-objective nurse scheduling problem: A case study of private hospital in Thailand

Nontawat Sukonpat, Pisit Jarumaneeroj — Wed, 20 Dec 2023 00:00:00 +0700

Nurse scheduling is one of the most important tasks that greatly affects the performance of nurses. Ineffective scheduling may lead to health and mental problems – caused by lack of adequate rest and work stress – as well as burnout and job quitting. To avoid a costly process of recruiting and selecting appropriate medical personnel, efficient nurse scheduling is therefore needed; and we achieve this by help of a mixed-integer linear programming model. We expect that this approach would help reduce the planning time, as well as the errors pertaining to the current scheduling method, while increasing scheduling fairness, which is typically neglected in practice. In doing so, we have first collected all necessary data from experience nurses and, later, transformed them into a mathematical formulation model with three different objectives: (i) to reduce the excess cost of operating nurses, (ii) to increase workload equity, and (iii) to increase nurse’s satisfaction measured by reducing constraint violations. This model test was conducted through a small problem. Then, the results were used to validate the accuracy of both objectives and constraints in order to be a guideline for the construction of a nurse scheduling program in the future.

K – Means Clustering and Hill Climbing Algorithm to vehicle routing problem on Excel VBA

Chonraya Thayod, Kanyarat Mueabsri, Kanate Puntusavase, Kawinthorn Saicharoen — Wed, 20 Dec 2023 00:00:00 +0700

The purpose of this study is to present the transport routing using the hill climbing algorithm (HCA), compared with the simulated annealing algorithm (SA) by stratifying the customers by K-Means clustering to group the transport coordinates of customers. The cost of fuel is calculated by distance-based transportation costs; so, the transportation route with the shortest distance is needed for the calculation. As the experimental result, it was found that customers should be grouped into six groups. This transport routing cost is compared to the previous routing's cost, and it showed that the transportation distance was reduced from 6,144 kilometers to 5,928 kilometers, reducing the distance by 216 kilometers. This reduced the transportation costs from 5,805,888 baht per year to 5,510,016 baht per year, which was calculated to be 295,872 baht reduction per year, or 5.10 percent of cost reduction per year.

Using Simulations to Reduce Waiting Time at The Outpatient Department of Laem Chabang Hospital

Thanapan Kongtong, Apinut Kumparn, Akarawit Cheewatummanon — Wed, 20 Dec 2023 00:00:00 +0700

This article has created a simulation model of the outpatient department at Laem Chabang Hospital in Chonburi Province in order to reduce waiting times for patients. The analysis using the model revealed that the bottleneck process in the department is the Clinic Rooms 1, 2, and 3, with average waiting times of 74.53±0.85 minutes, 75.77±0.83 minutes, and 75.79±1.15 minutes, respectively. Therefore, the article proposes 11 improvement scenarios, where Scenarios 1 to 4 involve individual improvements, and Scenarios 1 to 4 are combined to create Scenarios 5 to 11. The analysis results indicate that the recommended scenario is Scenario 11, which involves adding one doctor to the Clinic Rooms, adding one OP29 medication preparation staff, and adjusting the patient entry rate by allowing patients with yellow cards to enter until 12:00 PM and scheduling patients with blue cards from 1:00 PM onwards. By implementing these changes, it is possible to reduce the average waiting time for patients with blue cards from 127.24±1.00 minutes to 27.98±0.69 minutes, for patients with pink cards from 13.64±0.87 minutes to 13.61±0.82 minutes, for patients with white cards from 75.16±2.36 minutes to 70.95±1.87 minutes, and for patients with yellow cards from 163.48±1.54 minutes to 71.32±0.82 minutes.

Development of Mathematical Model for Humanitarian Logistics Management: A Case Study of Floods Disaster

Chanipa Uthanipan, Rojanee Homchalee — Wed, 20 Dec 2023 00:00:00 +0700

This research is development of mathematical model for humanitarian logistics management for a case study of floods disaster in Thailand, using Integer Non-linear Programming (INLP) to determine the location of shelters and the amount of sufficient resources to help the victims. The developed model has been applied to case studies in flood-affected arear, Kantharawichai District, Maha Sarakham Province, which has a total area of 10 sub-districts, 183 villages and has a total of 82,764 victims. There are three villages, of which 570 victims are unable to live in the villages had to be moved to a shelter. The research specifies the scope or major limitations of the problem as follows: 1) there are two relief distribution centers, namely the Maha Sarakham Provincial Disaster Prevention and Mitigation Office and the Kantharawichai District Office, 2) there are 19 alternative shelter centers, which are in various government areas such as schools and temples, and 3) transport of victims and relief items will be used only by trucks. The proposed INLP model was processed with the LINGO 13.0 program. The results showed that two shelters should be opened. The Chumchon Makha School Center is opened for receiving the 164 victims from Kut Hua Chang Village, and the Ban Makok School Center is opened for receiving the 190 victims from Ban Khong Kut Wien, and 216 victims from Ban Huai Chan. For the 180 flood-affected villages, most of them were allocated relief items from Kantharawichai District Office which has a distance closer than from the Maha Sarakham Provincial Disaster Prevention and Mitigation Office. Moreover, both the transportation of the victims and the distribution of relief items should be transported by a 4-wheel operating truck. In addition, the solutions of proposed INLP model also showed that all victims are cover helped, with the minimize total operating cost of 20.859 million THB. Furthermore, the developed model can also be applied to larger scale flood disaster management as a result of the sensitivity analysis, and it can also be further developed for humanitarian logistics management in case of other types of disasters.