STOCHASTIC OPTIMIZATION TECHNIQUES FOR MODELING UNCERTAINTY IN EPIDEMIOLOGICAL FORECASTING

Accurate forecasting amidst health uncertainties is more critical than ever, especially in countries like Ghana where disease volatility challenges traditional prediction models. This study addresses the urgent need for resilient forecasting systems by applying stochastic optimization techniques to epidemiological modeling. Focusing on stochastic programming, metaheuristic algorithms, and robust optimization methods, the research aimed to enhance the accuracy and adaptability of health forecasts. Using secondary data from Ghana Health Service, Ministry of Health, and WHO reports between 2020 and 2024, statistical analyses including correlation and regression were conducted. Major findings revealed that metaheuristic algorithms exhibited the strongest positive influence on forecasting accuracy (r = 0.754, p < 0.01), and stochastic programming models significantly reduced prediction errors by 42%, while robust optimization techniques minimized worst-case deviations by 51%. The overall model explained 65.7% of the variance in forecasting accuracy, confirming a strong relationship (R² = 0.657). These results underscore that integrating stochastic methods substantially enhances public health forecasting, paving the way for more effective interventions. Consequently, the study recommends broader adoption of stochastic frameworks, improved data infrastructures, and targeted training programs to embed these methodologies into Ghana’s national health strategy.