A Bilevel GA–PSO–MILP Framework for Capacity Planning and Day-Ahead Scheduling of an Integrated Electricity–Heat–Cooling Energy System

Abstract

This paper proposes a bilevel planning–operation framework for an integrated electricity–heat– cooling energy system comprising photovoltaic generation, wind generation, a gas-turbine-based combined heat and power unit, a ground source heat pump, a gas boiler, battery storage, thermal storage,  an absorption chiller,  and an electric chiller.   The  upper  level determines the installed capacities of major devices, while the lower level performs day-ahead coordinated dispatch over four representative days.  To preserve the engineering logic of the original optimization program while improving analytical transparency and reproducibility, the planning layer is formulated as a GA–PSO-based search procedure and the coupled planning–operation problem is further expressed as an equivalent mixed-integer linear benchmark model.  A complete mathematical formulation is provided, including explicit decision-variable definitions, energy-conversion relationships, storage dynamics, logical constraints, and a comprehensive notation table.  Using the representative-day data embedded in the program package, the optimal configuration is obtained as 800 kWh battery storage, 900 kW photovoltaic, 600 kW wind, 1200 kW gas turbine, 300 kW ground source heat pump, 200 kW gas boiler, 600 kWh thermal storage, 479.08 kW absorption chiller, and 150 kW electric chiller. The corresponding daily equivalent annualized investment cost is 6286.79, the day- ahead composite operating objective is 77400.43, and the ex-post composite operating objective under realized loads is 78763.31.  The results indicate that strong renewable deployment, flexible storage scheduling, and coordinated electricity–heat–cooling conversion can substantially improve the economic and operational performance of integrated energy systems.