The choice of algorithms for solving a multi-agent routing problem based on solving related problems

The paper considers the problem of reducing the complexity of $NP$-hard problems by using related problems for which an optimal or acceptable solution is already known. For multi-agent routing tasks, a technique is used based on network clustering consistent with traveling salesman routes on each cluster and constructing routes that take into account the limitation of delivery time windows. A mathematical model is presented that corresponds to a block of pseudo-Boolean conditional optimization with constraints in the form of disjunctive normal forms that allows polynomial solvability and a block of time constraints. The results of choosing metaheuristics based on related problems are used in a program for the delivery of goods by many agents to consumers located at the vertices of the regional infrastructure road network.