An Approach to Knowledge Representation for Vehicle Routing Problems

Abstract

Aiming to overcome deficiencies of existing DSS for VRP, which cannot meet with real-time and online routing and scheduling, a new knowledge representation approach to efficiently eliciting domain knowledge of the vehicle routing problems (VRP) and transforming this elicited knowledge into a machine usable format is proposed. According to the characteristics of VRP, the approach contains three steps: confirming problem rarr identifying problem structure rarr representing problem knowledge. The approach is demonstrated in detail through case study after analyzing principles of it. This approach sets foundations for consequent modeling and solving of VRP and can be used as reference by DSS for any types of VRP