Dynamic pricing based on model-predictive management in marketplace ecosystems
Artem Yu. Varnukhov, Assistant, Business Informatics Department. Ural State University
of Economics
Abstract
This article proposes an approach to pricing based on the concept of model-based predictive management, in which the pricing process is viewed as an optimization problem, with price acting as the control variable. The proposed approach is based on a combination of predictive forecasting, a decision-making system, and a genetic algorithm. The predictive model is used to forecast future demand scenarios generated by simulated pricing decisions and market environment parameters. The decision-making system uses these forecasts to evaluate alternative price trajectories subject to specified constraints. A genetic algorithm is used to find the optimal trajectory that maximizes the specified objective function over the planning horizon. The developed approach was tested in a controlled environment, where it was compared with basic pricing strategies: a fixed-price strategy and a heuristic strategy based on specified rules. The results of the experiment showed that the proposed approach allows for higher total profit and revenue, increased demand satisfaction, a reduction in lost sales, and the elimination of out-of-stock situations.
Keywords: pricing; price optimization; marketplaces; model-predictive management; machine learning methods; evolutionary algorithms; price trajectory modeling.
For citation: Varnukhov A. Yu. Dynamic pricing based on model-predictive management in marketplace ecosystems. Digital Models and Solutions. 2026. Vol. 5, no. 2, pp. 36–54. DOI: 10.29141/2949-477X-2026-5-2-3. EDN: GEMVRP
