A bi-level, many-objective stochastic optimization model for handling simultaneous uncertainties in the operation of multi-carrier integrated energy systems: A hospital case study

• Developing a stochastic model using IGDT to handle many uncertainties in MIESs. • Considering risk-averse and risk-seeking decision policies in risk management. • Implementing a bi-level, many-objective model to capture system interdependencies. • Using a robust NSGA-III algorithm to solve the bi-level, many-objective model. The operating multi-carrier integrated energy systems (MIESs) at the building level face challenges from multiple deep uncertainties related to resources, demands, and market prices. These uncertainties complicate decision-making and affect system reliability, cost-efficiency, and sustainability. Tackling these challenges requires advanced optimization models capable of handling many simultaneous uncertainties. This study addresses this need by developing a bi-level, many-objective stochastic optimization model. The upper-level optimization applies information-gap decision theory to evaluate risk-averse and risk-seeking decision policies, ensuring the system’s adaptability under many uncertain parameters. This level uses a non-dominated sorting genetic algorithm III to solve the many-objective optimization problem, generating a many-dimensional Pareto efficient solution set. It then utilizes a hybrid decision-making tool that combines the fuzzy satisfying method with a conservative approach to select the final optimal solution from this set. The lower-level optimization formulates a mixed-integer nonlinear programming problem to minimize the costs of energy, emissions, and multi-carrier energy not supplied while maintaining technical constraints. The model validates its effectiveness across diverse real-world urban settings, including hospital and industrial applications, demonstrating its ability to generate robust operational strategies for MIESs. The results confirm that the proposed model outperforms conventional methods by providing efficient operational strategies for MIESs under uncertainties.