This paper examines efficiencies achieved by optimal scheduling of power generation equipment and electrical loads in a hybrid smart micro-grid under conditions where load activity may be delayed or advanced with negligible impact to system performance. We model military expeditionary energy systems to analyze the performance of existing unmanaged components and estimate potential savings obtained by coordinated management of battlefield heating and cooling systems. We employ rolling horizon optimization models to examine performance under varying degrees of uncertainty about future load demand and renewable production. We propose a novel mechanism to reduce power production costs through optimal prescriptive scheduling of loads, reducing peak demand and generator peak-to-average power ratios and facilitating a persistent shift to higher fuel efficiencies. In contrast to existing methods that employ either supply-side or demand-side management, we propose intelligently coordinating both sides to achieve greater efficiency. Using sensitivity analysis, this paper quantitatively demonstrates how grid composition, temperature band tolerance, and energy storage capabilities contribute to fuel efficiency under this approach.