ARDL Models Optimize Retail Sales Forecasting and Fair Food Pricing.

Researchers have investigated a method for optimizing retail sales forecasting and pricing strategies, particularly for food products, by integrating consumer fairness considerations. The study utilizes a log-log Autoregressive Distributed Lag (ARDL) model to predict total retail trade sales, where the price coefficient represents sales elasticity. The core innovation lies in framing the pricing problem as maximizing forecast sales while adhering to price ceilings tied to the Consumer Price Index (CPI), thereby safeguarding against consumer exploitation. The problem is solved using both Linear Programming and Simulated Annealing, applied to single and multi-product scenarios. A notable finding was that nominal elasticities appeared positive, suggesting an unconstrained sales maximizer would push prices to their upper limits, making the CPI ceiling crucial for fairness. Simulated Annealing, however, yielded more conservative, interior prices that balanced sales targets with lower consumer costs. The framework offers a transparent, fairness-aware approach to dynamic pricing.