A Dynamic Simulation/Optimization Model for Scheduling Restoration of Degraded Military Training Lands.

A serious challenge to natural resource managers on Army installations is how to optimally maintain safety and training realism on maneuver lands using limited funding.  We will report on the development of a rolling horizon approach that seeks to address the optimal scheduling of land restoration for a given long-term training schedule and availability of man power, equipment, and financial resources using a dynamic simulation-optimization framework. Intensive use of military tracked vehicle maneuver lands causes deterioration in ground surface quality (loss of vegetation and tracks caused by vehicles), increases the likelihood of damages in stream crossings, and facilitates the formation of gullies when followed by heavy rain. Inadequate restoration of the degraded lands, due to resource limitation and/or improper scheduling of restoration efforts, may restrict the scheduled training activities and jeopardize personnel and equipment safety.

Using a discrete dynamic optimization model, we determine the optimal scheduling of land repairs over time and space to minimize the adverse effects on the training landscape, taking into account short and long-term forecasted weather, scheduled maneuver exercises, and unique qualities and importance of the maneuver areas. Since short-term weather forecasts are more reliable in the optimization model we use a higher weight for short-term benefits/damages and also assign a spatially varying weight factor to individual training land parcels (a higher weight assigned to more critical training areas). We move the scheduling period using a rolling horizon approach that optimizes the spatial/temporal land restoration schedule. The model is intended to inform short term allocation of land rehabilitation resources and then be rerun to consider new weather forecasts and actual damage repair.

The modeling framework is generic and applicable to any military installation. In this paper, we present the results of an application to Ft. Riley, KS. We divide the entire landscape into 78 training areas and determine the weekly optimal scheduling throughout a year. We perform a sensitivity analysis and generate optimal land restoration schedules and resulting impacts on the landscape quality under various resource limitations.