An interval rough improved ordinal priority approach-based decision support system to redesign postal and logistics networks

Abstract

Postal and logistics companies are essential subjects in the economy, providing services of the corresponding assortment for a wide range of business and private users. Service providers strive to meet the needs of users and, at the same time, make as much profit as possible. The efficiency of each of the subsystems in companies from this area significantly impacts the sustainability of postal and logistics systems. Rural areas, which are characterized by a smaller number of users and services and a low level of system efficiency, can have an additional negative impact on sustainability. As a result, optimization tasks become complex but also necessary to solve. The paper proposes an interval rough improved Ordinal Priority Approach - Power Schweiyer-Sklar Combined Compromise Solution (I-OPA - PSS'CoCoSo) methodology for prioritizing different models of solving the problem of inefficient network units. Methodological novelties are: a) A new approach for defining the lower and upper limits of interval rough numbers is proposed, which is based on nonlinear Bonferroni functions; b) The classic OPA linear model is improved through the implementation of a new concept for defining relational relationships between criteria; c) The CoCoSo method is improved through the implementation of nonlinear PSS and implementation of a novel function for the integration of aggregate strategies. The application of the interval rough I-OPA - SSP'CoCoSo methodology is demonstrated through a case study on the example of a public postal operator operating in the territory of the Republic of Serbia. Since this is a system with a highly developed infrastructure and network throughout the entire country, this further implies the applicability of the methodology to smaller systems or sectors within larger companies that deal with parcel deliveries and other logistics activities. A new aggregation function is introduced to define the compromise index of the alternatives as well as eliminate the anomaly of the original function. The simulation of different scenarios is enabled depending on the degree of risk. The proposed methodology enables decision-making in conditions of incomplete and imprecise criteria values. In accordance with the aforementioned, this approach contributes to improving the accuracy of modeling expert opinions, and consequently, in making the final decision.