Multi-CODP decision models for supplier selection and order allocation in customized logistics service supply chain

To address the problem of positioning customer order decoupling point (CODP) in customized logistics service supply chain and to provide logistics service integrators (LSI) with various CODP decision-making and supplier selection solutions. We consider LSI, logistics service providers (LSPs) and customers and propose multiobjective nonlinear team optimization models for multi-CODP location, including minimizing the cost to the LSI and maximizing the synthetic satisfaction of LSPs and customers. This paper adopts an improved nondominated sorting genetic algorithm (NSGA-II), in which the key modules such as the crossover operations, chromosome structure and mutation operations are reconstructed, to solve the team models and investigate the Pareto-optimal front for the two objectives. To demonstrate the validity of NSGA-II and the reliability of mathematical models, cases from a specific dataset are generated and solved. Moreover, a classic genetic algorithm (GA) is applied to solve for the two objectives separately. The optimal solutions found by GA are \(f_{1}\) $f_1$ = 2027.491, \(f_{2}\) $f_2$ = 0.7704 and \(f_{1}\) $f_1$ = 2454.512, \(f_{2}\) $f_2$ = 0.8909. The corresponding optimal solutions in the Pareto-optimal solution set are \(f_{1}\) $f_1$ = 2053.675, \(f_{2}\) $f_2$ = 0.7715 and \(f_{1}\) $f_1$ = 2427.114, \(f_{2}\) $f_2$ = 0.8843. We find that the optimal solutions provided by GA cannot dominate any solution in the Pareto-optimal solution set, indicating that the results are valid and sufficiently convergent. In summary, the results of this paper provide LSI with easy access to various supplier selection schemes with multi-CODP, which is essential to reduce the cognitive burden on decision makers and address the diverse customization needs of many challenging real-world logistics issues.