مدل‌سازي و بهينه‌سازي مسئله مسيريابي ناوگان سرويس كارمندان در شركت‌هاي توليدي و خدماتي

Modeling an‎d Optimization of Employee Shuttle Fleet in Manufacturing an‎d Service Companies

مهندسي صنايع و آينده پژوهي

Vehicle routing is one of the most widely applied problems in industrial engineering an‎d logistics, aiming to develop an optimal plan for service delivery with minimum cost an‎d time. The transportation of employees plays a significant role due to its impact on organizational productivity. This research investigates the challenges associated with managing employee transportation fleets. To address this problem, a comprehensive mathematical model is developed with the objective of minimizing total costs. Unlike most existing studies that primarily focus on freight transportation, this research considers the specific conditions an‎d constraints of human transportation an‎d models the problem in two distinct yet interrelated phases. In the first phase (return mode), the process begins at the company location, where employees board the buses, an‎d the buses distribute them across various dro‎p-off points. In the second phase (outbound mode), the process is reversed: buses start from multiple points, pick up employees, an‎d head toward the company. To ensure alignment with real operational conditions an‎d prevent deman‎d from exceeding available capacity, the total number of trips operated from each depot must not exceed the number of buses stationed there. To solve the proposed model an‎d assess its performance, a combination of exact an‎d metaheuristic methods is employed. For small-scale (10–28 nodes) an‎d medium-scale (30–50 nodes) instances, GAMS software is used to obtain optimal solutions. Considering the NP-hard nature of the problem an‎d increasing complexity in large-scale instances (70–120 nodes), three metaheuristic algorithms—Genetic Algorithm (GA), Artificial Rabbit Optimization (ARO), an‎d Harris Hawks Optimization (HHO)—are implemented. Key contributions of this research include the detailed two-phase modeling of outbound an‎d return routes, incorporating delay an‎d time modeling, an‎d employing novel an‎d efficient optimization algorithms in this domain. The results indicate that as the problem size increases, computational complexity an‎d solution time grow significantly, highlighting the necessity of intelligent solution methods. Among the eva‎luated algorithms, ARO demonstrated superior performance in both phases, providing the lowest average objective function value—which represents reduced transportation costs an‎d computation time—compared to GA an‎d HHO. Quantitative comparisons revealed that although GA performed satisfactorily an‎d close to ARO in small an‎d medium scales, its computation time increased sharply for larger problem sizes. Conversely, HHO consistently produced the highest objective values an‎d the weakest performance across all problem dimensions. In addition to solving the model, a detailed sensitivity analysis was conducted on the cost per kilometer an‎d average travel time per kilometer parameters. The analysis showed that the proposed model is highly sensitive to cost variations. For example, in the outbound phase, increasing the cost coefficient per kilometer from 0.5 to 1.5 raised the objective value from 6.956 to 8.453, an‎d increasing the average travel time coefficient per kilometer increased the objective value from 2.328 to 12.222. Similarly, in the return phase, increasing the cost coefficient from 0.5 to 1.5 raised the objective value from 5,421 to 6.567, an‎d increasing the average travel time coefficient raised it from 1.801 to 11.191. Overall, the outcomes of this study can serve as a practical solution for improving the efficiency of organizational urban transportation systems, significantly reducing fleet costs, an‎d ultimately enhancing employee satisfaction an‎d productivity in manufacturing an‎d service industries.

5

156690