To Start Up a Start-Up$-$Embedding Strategic Demand Development in Operational On-Demand Fulfillment via Reinforcement Learning with Information Shaping

2025-04-08

AI Summary

On-demand transportation startups face cold-start challenges in new markets, where limited initial supply constrains demand growth. Method: We propose an “information-shaping” mechanism that models long-term regional demand growth as a dynamic feedback process governed by real-time vehicle dispatch decisions. Our approach integrates closed-form convex optimization solutions to guide reinforcement learning data generation, designs a state-action value decomposition variant of DQN, and incorporates counterfactual data augmentation. Contribution/Results: The core innovation lies in endogenously aligning short-term dispatch policies with sustainable demand cultivation. Extensive multi-city simulations demonstrate that our method increases cumulative demand volume by 27.4% over 3–5 years and improves user retention in key regions by 19.8%, significantly outperforming pure efficiency-oriented baselines.

Technology Category

Search and Optimization: Learning to SearchPlanning, Routing, and Scheduling: Planning with Language ModelsApplication Domains: Transportation

Application Category

Economics, Online Markets and Human Computation: The sharing economySearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingUser Modeling, Personalization and Recommendation: User privacy protection in personalized systems

Abstract

The last few years have witnessed rapid growth in the on-demand delivery market, with many start-ups entering the field. However, not all of these start-ups have succeeded due to various reasons, among others, not being able to establish a large enough customer base. In this paper, we address this problem that many on-demand transportation start-ups face: how to establish themselves in a new market. When starting, such companies often have limited fleet resources to serve demand across a city. Depending on the use of the fleet, varying service quality is observed in different areas of the city, and in turn, the service quality impacts the respective growth of demand in each area. Thus, operational fulfillment decisions drive the longer-term demand development. To integrate strategic demand development into real-time fulfillment operations, we propose a two-step approach. First, we derive analytical insights into optimal allocation decisions for a stylized problem. Second, we use these insights to shape the training data of a reinforcement learning strategy for operational real-time fulfillment. Our experiments demonstrate that combining operational efficiency with long-term strategic planning is highly advantageous. Further, we show that the careful shaping of training data is essential for the successful development of demand.

Problem

• Optimizing fleet allocation for on-demand start-ups in new markets
• Balancing operational efficiency with long-term demand growth
• Using reinforcement learning to shape strategic demand development

Innovation

• Reinforcement learning optimizes real-time fulfillment operations
• Strategic demand development integrates with operational decisions
• Training data shaping enhances demand growth success