A Two-Stage Stochastic Mixed-Integer Programming Approach to the Smart House Scheduling Problem

SUMMARY A “smart house” is a highly energy-optimized house equipped with photovoltaic (PV) systems, electric battery systems, fuel cell (FC) cogeneration systems, electric vehicles (EVs), and so on. Smart houses are attracting much attention recently because of their enhanced ability to save energy by making full use of renewable energy and by achieving power grid stability despite an increased power draw for installed PV systems. Yet running a smart house's power system, with its multiple power sources and power storages, is no simple task. In this paper, we consider the problem of power scheduling for a smart house with a PV system, an FC cogeneration system, and an EV. We formulate the problem as a mixed-integer programming problem, and then extend it to a stochastic programming problem involving recourse costs to cope with uncertain electricity demand, heat demand, and PV power generation. Using our method, we seek to achieve the optimal power schedule running at the minimum expected operation cost. We present some results of numerical experiments with data on real-life demands and PV power generation to show the effectiveness of our method. © 2013 Wiley Periodicals, Inc. Electr Eng Jpn, 186(4): 48–58, 2014; Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/eej.22336