International Journal of Industrial Optimization

A heterogeneous fleet electric vehicle routing model with soft time windows

Yoanda Astri Ayu Kinanti — 2024-09-03

The emergence of electric vehicles in distribution and logistics activities has brought significant benefits due to their unique characteristics, such as energy-efficient and lower carbon emissions. In the perspective of vehicle routing problem, electric vehicles pose challenging constraints regarding the limited battery capacity, and thus their traveling ranges, and the availability of charging stations. In this paper, we propose a model of the fleet electric vehicle routing problem (EVRP) with soft time windows, where a mixed integer linear programming framework is implemented in model formulation. The objective of mathematical programming is to minimize the total operational cost, which consists of a fixed cost, a traveling cost, a battery charging cost, and probably a penalty cost due to time window violation. We implement our model in two simple cases, namely homogeneous and heterogeneous fleets EVRPs, characterized by loading and battery capacities. Each case consists of one depot, five customers, two electric vehicles, and two charging stations. Optimal routes are obtained using the well-known branch-and-bound method under Lingo 17.0. It is found that the existence of charging stations may affect the routes selection and the implementation of soft time windows rather than hard time windows has been proven to increase the feasibility of routing problem.

Optimizing welding parameters for high deposition efficiency in waam by using the taguchi method

Ahmad Baharuddin Abdullah — 2024-09-03

Wire arc additive manufacturing (WAAM) is a type of additive manufacturing technology that offers high flexibility in shaping products and is cost-effective due to its low material consumption and rapid time to market. Material consumption can be evaluated by assessing deposition efficiency during welding. The efficiency of a deposited metal depends on various processes and welding parameters, including travel speed, wire feed rate, voltage, distance of the torch from the base, and many others. Therefore, process capability can be efficiently achieved by crucially determining the key parameters that have the most significant effect. In this study, the main objective is to determine the most significant parameters to obtain the optimum deposition efficiency of a gas metal arc welding-based 3D welding machine. The Taguchi experimental design method is used to determine the optimal welding parameters. Results showed that the distance of the torch from the base is the most significant parameter, followed by welding speed and wire feed rate. The observation is validated via a confirmation test.

Development of genetic algorithm for human-robot collaboration assembly line design

Anas Ma'ruf — 2024-09-03

An assembly line requires flexibility due to a shorter product life cycle. A way to increase flexibility is to utilize collaborative robots or cobots. Due to frequent product changes, redesigning an assembly line requires an efficient algorithm. This research aims to develop a genetic algorithm (GA) for solving a human-cobots assembly line design. The setup time of cobots is considered due to the flexibility of conducting multiple tasks by exchanging tools / end-effectors. The main contribution of the research is the efficient GA for solving assembly lines considering setup time. Secondly, the study proposed an upper limit parameter that enables faster computation without sacrificing the quality of the solution. The computational results showed that the algorithm could achieve an optimal solution with the number of tasks less than 35. Experiments of several data prove the proposed GA obtained solutions with an average gap of 3.83% to the optimal solution. Also, a faster computation time with an average difference of 64.66%. The proposed GA obtained a reasonable solution with fast computing time that helps improve efficiency and effectiveness in decision-making related to frequent redesigning of assembly lines.

Implementation of 5s and kaızen methods for developing a novel wage assessment method in a steel construction factory: an application in Turkey

Mehmet Burçin Önay — 2024-09-03

The study aims to implement the 5S (seiri, seiton, seiso, seiketsu, and shitsuke) method and KAIZEN for emphasizing the troubles and defective products, establishing work standards, implementing fair wage assessments based on job analysis and job evaluations in a steel construction factory. A more objective wage assessment method is developed, and workers' unrest can be resolved fairly. 5S and KAIZEN studies have been applied for two years in a steel construction factory. Then, the evaluation of success factors within the internal structure of wage brackets utilized last year's 5S scores to enhance employees' confidence in the objectiveness of the wage system assessment. A reformer method for assessing wages has been created and implemented, integrating lean manufacturing principles and a job analysis and evaluation system. The framework has been tested and implemented only for a steel construction factory. In the future, studies could be conducted to assess different sector factories. The proposed framework has been successfully implemented in a medium-large scale manufacturing factory. A novel wage assessment framework that involves lean application studies integrated into the job evaluation method has been proposed in a medium-sized manufacturing factory.

Product pricing based on customer perception quality and service convenience using interval type-2 fuzzy logic system

Muhammad Ridwan Andi Purnomo — 2024-09-03

In the competitive landscape of customer goods, particularly in the wrapping paper industry, pricing strategies are critical to achieving market success. This study presents a novel approach to product pricing by integrating customer perception quality and service and convenience factors using interval type-2 fuzzy logic system (IT2FLS). The customer perception quality factor is subdivided into material quality and aesthetics design sub-factors while the service and convenience factor comprise web-based ordering system as well as the web-based post-sale customer engagement. The methodology involves collecting data through customer surveys and expert evaluations to quantify the perceived importance and performance of each sub-factor. The IT2FLS is employed to handle the inherent uncertainty and imprecision in experts’ judgment, providing a robust framework for aggregating these qualitative assessments into a comprehensive pricing model. This IT2FLS allows for the dynamic adjustment of pricing based on fluctuating customer perceptions and service levels. The outcome of the proposed IT2FLS is a pricing factor that serves as a multiplier for the standard product price established by the company. The new product prices have been validated also considering historical data and it was found that the prices remain acceptable to customers without drastically impacting sales. This study contributes to the body of knowledge on pricing strategies by offering a sophisticated, mathematically grounded approach that accounts for the complex, fuzzy nature of customer preferences. The proposed model not only enhances pricing accuracy but also provides a flexible tool for managers to adapt pricing strategies in real-time based on customer feedback and service performance.

Prediction analysis of retail store sales level using neural network algorithm method based on customer segments

Mylenia Martina Yuniar — 2024-09-03

Marketing activities are of significant importance to business operations, as they are uniquely positioned to provide value to consumers. The marketing mix represents one of the strategic approaches employed to attain these organizational objectives. However, the company's sales data is only available for consultation in the archives. By understanding customer preferences and requirements, the company can readily develop an effective marketing strategy to compete with similar businesses. Accordingly, this study employs the neural network methodology to forecast sales based on the company's historical sales data. The research method employs a neural network due to its capacity for processing substantial data sets with flexibility. Moreover, the Root Mean Square Error (RMSE) must be employed to ascertain the precision of the utilized model. The findings of this study indicate that the discrepancy between the actual and predicted values is minimal, suggesting that the model is able to accurately represent the data. Similarly, the results of the RMSE (Root Mean Square Error) demonstrate that the model's accuracy is improving, with minimal values observed in each segment. A 4P marketing mix strategy may be employed to enhance the company's sales potential. Based on the findings of the research, it can be posited that the results of the prediction data set, the visual prediction results, and the RMSE using the Neural Network method can be utilized effectively and accurately to forecast sales and assist company owners and management in considering target sales levels in the future.