Artificial Intelligence and the Management Science Practitioner: Knowledge Enhancements to a Decision Support System for Vehicle Routing

The application reported in this paper is a working prototype of a new generation of DSSs that more fully support decision makers. Like many DSSs it uses a general algorithm to solve an abstraction of the real problem. Because only an abstraction was solved, a KBES was used to apply knowledge of the general domain to move to a solution that is feasible and satisfactory for the real problem. This role for a KBES appears to have wide applicability in a setting where a model is insufficient by itself. In addition to demonstrating this role, we have identified two other generalizable roles: model selection and sensitivity analysis. We expect DSSs of the future to use KBESs in all three roles.