DWARP builds on the WARP optimizing aircraft router developed by On Time Systems for the U.S. Air Force’s Air Mobility Command. DWARP allows many mission routings to be jointly optimized taking into account the limited availability of resources such as airspace clearances, aircraft parking space, and time, resulting in a minimum-cost/maximum-value overall assignment of resources.
While finding optimal routes for individual aircraft is a complex search problem that requires sophisticated algorithms, managing the interactions among multiple sorties and constraints requires the ability to search an even larger space and hence even more powerful techniques. For example, sorties may be advanced or delayed by varying amounts, rerouted around restricted airspace, or perhaps flown at different altitudes. Since each different choice may restrict the choices available to other sorties, the problem can quickly become combinatorial in the number of sorties.
DWARP searches in a two-tiered fashion, partitioning the search space into problems that involve only one aircraft and those that require information about several planes to solve. DWARP breaks the problem down by distinguishing between searching for the optimal distribution of resources across multiple missions and searching for the optimal individual mission routing given these constraints.
Given limited resources, it may be impossible to find optimal solutions for all aircraft, so tradeoffs must be made to avoid degrading the global solution unnecessarily. The second tier of DWARP search works at this level, by attempting to find the allocations of resources to each mission that result in an optimal overall plan. Based on the routing solutions provided by WARP, DWARP attempts to distribute resources in a way that maximizes overall performance, according to AMC’s metrics.
DWARP was incorporated into the GLOBE system developed for the USAF's Global Response and Synchronization Advanced Technology Demonstration.