Fragmentary Models for Some Classes of Vehicle Routing Problems
Abstract
Most of the optimization problems related to the delivery of goods belong to the class of nondeterministic polynomial time hard (NPH) problems. For such problems, the use of various metaheuristics is justified. However, the presence of a large number of constraints, first, significantly complicates the construction of the corresponding mathematical model, and, second, makes the method of solving the problem individual, applicable to a rather narrow class of applied problems. The authors have shown that a fragmentary model can be built for many classes of logistics problems. This allows them to reduce the optimization problem of the corresponding class to an unconstrained optimization problem on a set of permutations. The authors propose a method for constructing universal hybrid algorithms for finding suboptimal solutions for certain classes of vehicle routing problems based on a combination of fragmentary and well-known metaheuristics.