Improving Routing Optimization Algorithms: A Business Rules Based System to Generate Initial Solutions

Business rules serve as a foundational concept in computational problem-solving, offering a structured approach to encoding knowledge and defining actions within a system. These rules, reminiscent of logical statements, consist of conditions (if-part) and corresponding actions (then-part), encapsulating the logic governing decision-making processes. Importantly, they also enable the modelling of human reasoning, allowing systems to emulate cognitive processes and make decisions akin to those of human operators. This aspect adds a layer of cognitive fidelity to computational systems, enhancing their ability to perform complex tasks and adapt to diverse problem domains. In this work, the authors propose a hybrid approach that integrates a system of business rules to define the initial solution in an optimization process. In particular, the analysis of the impact of using this initial feasible solution in a Tabu Search algorithm for solving a vehicle routing problem with time windows is presented. For this purpose, a case study corresponding to a real freight distribution problem in Catalonia (Spain) is implemented to analyse if the obtained results outperformed other typical algorithms for generating the initial feasible solution of this type of routing algorithm.