Optimal Design of Energy Systems Using Constrained Grey-Box Multi-Objective Optimization.

Area of Science:

• Computational Science
• Optimization Theory
• Energy Systems Engineering

Background:

• Global optimization of complex energy systems is challenging due to noisy derivatives and computational expense.
• Multi-objective problems require generating trade-off Pareto-optimal solutions.
• The p-ARGONAUT system was previously developed for single-objective grey-box optimization.

Purpose of the Study:

• Extend the p-ARGONAUT system to address multi-objective optimization problems.
• Evaluate the performance (accuracy, consistency) of the extended framework under numerous equality constraints.
• Compare the framework's performance against other derivative-free optimization solvers.

Main Methods:

• Adaptation of the p-ARGONAUT (parallel AlgoRithms for Global Optimization of coNstrAined grey-box compUTational problems) system for multi-objective problems.
• Development of accurate and tractable surrogate formulations for unknown equations.
• Testing on benchmark multi-objective problems and a commercial building energy market design case study.

Main Results:

• The extended p-ARGONAUT framework demonstrates effective performance in multi-objective optimization.
• The system shows accuracy and consistency, even with multiple equality constraints.
• Comparative analysis indicates competitive performance against existing derivative-free solvers.

Conclusions:

• The extended p-ARGONAUT system provides a robust and efficient approach for multi-objective global optimization of complex systems.
• The framework successfully handles challenging energy system optimization problems with multiple objectives and constraints.
• This work advances the capability of derivative-free optimization for energy systems and related fields.