A dual approach to optimization in wastewater treatment plants: Deterministic and stochastic perspectives

Efficient design of wastewater treatment plants (WWTPs), particularly in terms of biological process configuration and operational cost optimization under variable conditions, is critical for sustainable performance. This study compares a deterministic optimization approach, which assumes fixed input conditions, with a stochastic framework that incorporates uncertainty through Monte Carlo simulations and metaheuristic algorithms—Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). A municipal WWTP was modeled under three configurations (Conventional, Ludzack-Ettinger, and Bardenpho) to evaluate operational costs and effluent quality indicators, including COD (≤90 mg/L), BOD (≤30 mg/L), and NH₃-N (≤15 mg/L), under uncertainty. Stochastic optimization achieved up to 68.5 % cost reduction and ensured effluent compliance, with PSO emerging as the most cost-effective method. Sensitivity analysis revealed hydraulic retention time and dissolved oxygen as the most influential variables affecting performance. Beyond cost savings, the novelty of this work lies in integrating uncertainty analysis with metaheuristic optimization across multiple plant configurations, providing a robust, scenario-based decision-making tool for real-world WWTP management. • Deterministic and stochastic approaches to the optimization of wastewater treatment plants are compared. • The model incorporates uncertainty in key parameters such as hydraulic retention time and dissolved oxygen concentration. • Stochastic optimization achieves a total cost reduction of up to 68.5 %. • Particle swarm optimization and genetic algorithm are the most efficient and consistent methods for minimizing total annual costs. • The Bardenpho process achieves the highest total nitrogen removal.