Iterative on-line solution of minimum-time/limited-fuel/fixed-route problems

Operations research has generated many approaches to solve the optimal path or route assignment problem. The authors discuss a technique for rapidly finding the optimal traversal strategy after the route itself has been assigned: nonlinear programming via a modification of Newton's algorithm. A key to the speed of computation is the use of reduced-order modeling of the traversal process. The resultant rapid solution allows for online updating of strategies in the face of modeling uncertainties and potential changes in downstream operating conditions. Frequent reevaluation and updating of the optimal strategy affects a global strategic feedback over the entire history of the route traversal process and thus improves the robustness of the optimization routine. Solar/electric vehicle operation is considered as an example.<
>