Dynamic pricing strategy design for manufacturing service providers in manufacturing platforms

Existing game theory and scheduling optimization-based dynamic pricing methods suffer from complex decision-making processes, failing to adapt to service providers' (SPs) dynamic pricing needs in manufacturing service recommendation scenarios. To address this gap, this study proposes four rule-based dynamic pricing strategies—cost-plus pricing (CPP), service recommendation outcome-driven pricing (SROP), service capacity surplus-driven pricing (SCSP), and market average price-driven pricing (MAPP)—by analyzing SPs' dynamic response mechanisms during service recommendation. Additionally, integrating SPs' adaptive learning behaviors in competitive environments, two reinforcement learning (RL)-driven adaptive dynamic pricing methods were developed. These six strategies were embedded into a multi-agent simulation model abstracted from a real service recommendation system to evaluate their performance across diverse market environments. Results show that: (1) Among the rule-based strategies, SROP exhibits superior competitiveness in most scenarios due to its direct linkage with recommendation outcomes; (2) RL-driven pricing methods do not consistently outperform their rule-based counterparts, indicating that one-sided pursuit of dynamic learning capabilities may not help SPs establish market advantages. This study provides actionable pricing references for SPs in manufacturing service platforms and enriches the theoretical framework of dynamic pricing in service recommendation contexts.