College of Business of Shanghai University of Finance and Economics recently hosted a successful Seminar on Platform Operations, co-organized by the Department of Operations Management and the Shanghai Research Base for Data Science and Decision-Making Frontiers. The event gathered distinguished scholars, who engaged in in-depth discussions and vibrant exchanges on platform operation-related topics, delivering a rich and insightful academic experience.

The seminar commenced with a keynote address by Professor Wei Hang, Dean of the SUFE Graduate School. He emphasized that, as digitalization continues to sweep across the globe, the platform economy has emerged as a vital engine of economic growth. This seminar, with its focus on platform operations and its assembly of distinguished scholars, offers a valuable forum for advancing academic research and practical innovation. Professor Wei expressed his hope that the gathering would spark intellectual synergy and yield meaningful insights to support the sustainable development of the platform economy.

In his address, Professor Gao Weihe, COB Associate Dean, emphasized that the college has always been committed to cultivating business talent that meets the demands of the times. As a rapidly emerging and pivotal field, platform operations are closely integrated with the school's teaching and research initiatives. He highlighted the importance of the seminar in bringing together leading experts from various fields to present cutting-edge research findings. Professor Gao encouraged participating faculty and students to seize this valuable opportunity for in-depth learning and discussions, incorporating the insights gained from the event into their future academic endeavors and professional growth, ultimately elevating the college’s research in platform operations to new heights.

The first presentation was delivered by Professor Lu Xianghua from Fudan University, who shared his paper titled “Algorithm Advancement and Workforce Dynamics: Evidence from the On-Demand Delivery Market.” This research delved into the impact of AI-driven technologies on the gig economy labor market. Utilizing large-scale panel data of delivery riders, Professor Lu applied structural modeling and counterfactual simulations to systematically assess the effects of algorithmic improvements on labor participation, workforce composition, and the distribution of market performance. During the interactive session, participants engaged in in-depth discussions on topics such as algorithmic fairness, labor incentives, and skill upgrading mechanisms.

In his report titled "Research on the Dual Model System of 'Manufacturing Chain + Platform' and Its Value Co-creation Strategy," Professor Li Yongjian from Nankai University focused on the results of his research under a national key project regarding the "Manufacturing Chain + Platform" dual model. In addressing the common challenges in the manufacturing sector—such as supply-demand imbalances, entrenchment in the low end of the value chain, and intensified cross-industry competition from platform-based economies. Professor Li proposed that large manufacturing enterprises (such as Huawei and Apple) are extending downstream to build integrated ecosystems integrating hardware, software, and services (the "Manufacturing Chain + Platform" dual model) in order to achieve significant value enhancement (e.g., Apple's substantial platform contribution to its profits). This model contrasts with traditional single-product approaches. Through a comparison of the decline of Nokia (which focused heavily on products while neglecting platform development) with the success of Huawei and Apple (which effectively implemented the dual model), Professor Li emphasized the importance of understanding the fundamental differences between manufacturing chains (linear structures/product-based value) and platforms (network structures/ecosystem-based value). He then zoomed in on three core issues regarding integration: 1) How to overcome model differences to achieve "1+1>2" value co-creation; 2) Operational challenges in the various stages of platform development (user acquisition, retention, and value co-creation); 3) Strategies for third-party collaboration in hardware ecosystem expansion (independent vs. partnership, partner selection) and conditions for win-win outcomes. Professor Li concluded by stating that this model represents a critical pathway for upgrading the manufacturing industry, though both theoretical frameworks and practical applications still require further exploration.

Professor Feng Bo of Soochow University presented a report titled “Optimizing Incentives for Central Bank Digital Currency Adoption: The Role of Privacy and Network Effects.” The study focuses on strategies to promote the adoption of central bank digital currencies (CBDC), with a particular focus on the design of incentive mechanisms in light of privacy concerns and network effects. As an emerging payment instrument, the CBDC—specifically the Digital Currency Electronic Payment (DCEP) faces significant challenges, including a limited user base and intense market competition. By constructing a multi-homing two-sided market model, the research analyzes the asymmetric competitive dynamics between DCEP operators and third-party payment platforms (TPPs), and proposes an incentive framework grounded in consumers’ privacy preferences. The findings suggest that DCEP operators should adapt their incentive strategies flexibly, taking into account users’ sensitivity to privacy and the strength of network effects. When privacy is a primary concern for consumers, incentivizing merchants should take precedence; in markets with strong cross-network effects, reducing merchant commissions moderately can enhance competitiveness. The study further recommends that policymakers support the development of merchant infrastructure and continuously assess user transaction friction costs to facilitate broader DCEP adoption. During the interactive session, participants explored the comparative advantages of DCEP versus TPPs in terms of privacy protection, transaction costs, and processing speed, highlighting the potential of DCEP in decentralization and offline payment functionality. Overall, the research underscores that optimizing incentive strategies requires a nuanced understanding of market dynamics, user behavior patterns, and regulatory support in order to ensure the efficient rollout of DCEP and maximize consumer welfare.

Professor Shen Xiaobei from the University of Science and Technology of China delivered a presentation titled “Dynamic Pricing with Mental Accounting.” This study centers on the concept of mental accounting to examine the dynamics of supply and demand from the consumer’s perspective, and investigates how firms can formulate optimal dynamic pricing strategies when consumers behave irrationally. Professor Shen began by introducing the concept of mental accounting from the perspective of behavioral economics. She then developed a continuous-time theoretical model to analyze the associated optimization problems. The study rigorously derives the optimal pricing strategies for both homogeneous and heterogeneous consumer groups, offering robust theoretical validation. The findings provide a rational explanation for several complex pricing strategies observed in practice. These include anchor pricing with an initial price and penetration pricing accompanied by dynamic adjustments between high and low price segments. Overall, the research highlights how incorporating psychological factors into pricing models can lead to more effective and realistic pricing strategies in real-world applications.

Professor Liu Yan from the University of Science and Technology of China delivered a presentation titled “Dynamic Pricing of Limited Inventories with Word-of-Mouth Effect.” The talk explored the intersection of dynamic pricing and word-of-mouth (WOM) dynamics, focusing on how revenue management can be optimized under inventory constraints by leveraging consumer influence. Professor Liu began by introducing the concept and significance of the word-of-mouth effect, where consumers disseminate product information through social networks and media, thereby significantly shaping demand. He noted that traditional revenue management models often overlook this social influence. This research fills that gap by integrating WOM effects into dynamic pricing frameworks, particularly relevant in sectors such as hospitality, event planning, and fashion retail. He then detailed the study’s model, which deviates from conventional Poisson-based demand assumptions. Instead, it adopts a dynamic arrival process where WOM boosts future demand arrival rates. The pricing problem is formulated as a stochastic optimal control problem. However, due to the complexity of the Hawkes process used to model WOM-driven demand, analytical solutions are generally difficult to obtain. To address this, the study applies deterministic approximations to simplify the problem and derive optimal pricing strategies. In conclusion, the study offers both theoretical advancements and practical insights. It expands the role of WOM in revenue management and provides firms with a foundational framework to balance “exploration” (stimulating future demand through lower prices) and “exploitation” (maximizing immediate revenue). The presentation concluded with a lively discussion, where participants shared valuable perspectives on real-world applications and future research directions.

Associate Professor Long Zhenghua of Nanjing University, in his work "The Generalized c/μ Rule for Queues with Heterogeneous Server Pools," introduces the groundbreaking Generalized c/μ Rule (Gc/μ), marking a significant breakthrough in the optimization of queues within heterogeneous server pools. This research innovatively addresses the dual-objective optimization challenge under nonlinear cost structures inherent in the classic inverted V-shaped queuing model's dynamic routing problem. On one hand, it balances costs by employing a dynamic priority strategy that routes incoming customers at any given moment to the server pool with the minimal priority index, utilizing a priority function that reflects real-time system status. On the other hand, it ensures service level guarantees by incorporating a target allocation strategy that minimizes operational costs while adhering to predetermined service level constraints. The core breakthrough lies in extending the classical c/μ rule from linear cost frameworks to encompass general nonlinear holding and operational cost functions. Notably, the fluid approximation of the Gc/μ rule can degenerate into several well-known strategies, including fastest-server-first, load balancing, and idleness-based routing. The target allocation strategy dynamically adjusts routing decisions through state offsets without requiring prior knowledge of the system’s optimal solution, thereby markedly enhancing algorithmic adaptability. Finally, Professor Long’s simulations demonstrate that the Gc/μ rule effectively optimizes both holding and operational costs in heterogeneous resource allocation scenarios such as data center resource scheduling and cloud service platforms, providing a theoretical foundation and practical tool for the real-time optimization of complex queuing systems.

Professor Xie Lei from Shanghai University of Finance and Economics presented a report titled “Adaptive Learning and Ordering Strategies in Inventory Systems under Drifting Environments.” This study, grounded in empirical investigations of cross-border e-commerce enterprises, reveals that the primary challenges for businesses lie in traffic acquisition and bestseller product selection (as high-quality products often sell themselves), rather than in traditional revenue management models. This discrepancy highlights a significant gap between academic research and practical business needs. As a result, the research team pivoted toward a more realistic decision-making framework: inventory optimization in dynamically drifting environments. The study identifies two major limitations of conventional inventory models in real-world applications: first, drifting demand parameters (such as the non-stationary demand for perishable goods, rendering steady-state assumptions invalid); second, drifting cost parameter (for instance, real-time fluctuating logistics costs that distort profit calculations, since costs are not fixed). Professor Xie emphasizes that the study’s core innovation lies in simultaneously modeling the drift characteristics of both demand and cost parameters, thereby establishing a “dual-dynamic” decision environment that more accurately captures the uncertainties faced by enterprises. Regarding algorithm design for drifting environments, Professor Xie introduces a key improvement by eliminating abrupt change detection procedures—thus avoiding the complexity and data loss caused by traditional segmented history discarding. Instead, the approach employs a continuous learning mechanism that dynamically balances the “forgetting principle” (down-weighting outdated data to reduce bias) and the “information retention principle” (preserving some historical data to minimize estimation errors). This design avoids the need for complex detection modules and, as demonstrated in numerical experiments, outperforms traditional methods, offering both theoretical insight and practical value for inventory decision-making in volatile environments.

The seminar covered a wide range of topics, including gig economy platforms, manufacturing, digital currency, dynamic pricing, queue system optimization, and inventory management. The experts’ presentations combined profound theoretical insights with practical relevance, broadening the academic horizons of faculty, students, and participants of the College of Business of Shanghai University of Finance and Economics. Moreover, the seminar provided fresh perspectives and directions for advancing research and development in the realm of platform operations.