Comparing classical and metaheuristic methods to optimize multi-objective operation planning of district energy systems considering uncertainties

District energy systems (DES) can reduce CO2 emissions associated with buildings while meeting the energy needs of a group of buildings with fossil fuel or renewable energy resources that are located on-site. One of the present challenges of DES is optimizing the operation of energy components, as different optimization methods are available. These optimization methods can have various requirements for implementation, distinct needs for engineering labor, and may rely on freely accessible software or proprietary software. Most importantly, different methods may result in dissimilar operation planning for a given DES, which makes the selection of optimization method a key consideration for decision-makers. In this study, two optimization methods, a mixed-integer linear programming (MILP) solver as a classical method and a non-dominated sorting genetic algorithm II (NSGA-II) as a metaheuristic method, are used to optimize the early-stage operation planning of a hypothetical DES for a university campus in a cool and dry climate. The objective is to minimize the operating cost and CO2 emissions when considering uncertainties in energy demands, solar irradiance, wind speed, and annualized electricity-related emissions. Both methods present similar operation of energy components, operating cost, and operating CO2 emissions. The MILP solver and NSGA-II algorithm vary in computation time to perform the optimization, initial knowledge to run the simulation, accessibility (free/open-source status), and satisfaction of constraints. This work compares the characteristics of a MILP solver and NSGA-II algorithm to help future researchers select the suitable optimization method related to their case study. The software underlying this work is open-source and publicly available to be reused and customized for early-stage operation planning of their specific DES. This work is novel by optimizing the operation planning of a mixed-used DES to minimize the cost and CO2 emissions while considering uncertainties in weather parameters, energy demands, and annualized electricity-related emissions.