Framework speeds up rider incentives with minimal performance loss
A General Framework for Budgeted Threshold Incentives on Request
Artificial Intelligence
Summary
On-demand delivery companies pay riders extra bonuses based on recent rider activity, but running experiments to find the best bonuses can take months. The authors developed a flexible method that quickly predicts how different incentive plans will perform using short pilot data, speeding up decision making with very little loss in accuracy. Their method reuses past data to adjust predictions for new conditions, making it possible to respond quickly to changing budgets and rider groups. This approach greatly reduces the time and effort needed to design and adapt incentive plans while keeping rewards fair and effective.
What this means in practice
- •For ride-sharing platform operators: Adapt rider payment incentives quickly for holidays or weather without long randomized trials using faster budgeted planning.
- •For delivery logistics managers: Optimize on-demand delivery workforce bonuses under tight budgets with fast, data-driven planning that accounts for changing rider populations.
Authors
Zhuolin Wu, Chengrui Zhu, Wenhua Nie, Kenny Ye Liang, Junming Lin, Haiyang Li, Zhilin Li, Wenjia Geng, Zeyu Wu, Yinan Wu, Jinghua Hao, Renqing He
Abstract
On-demand delivery platforms pay riders through incentive activities whose tiers are set from recent completions of riders with a similar history. Operators request such plans for changing periods, rider populations, payment rules and budgets, often for holidays or bad weather, where randomized trials are scarce and take months to collect. We present a request-driven framework that composes four stages (conditional prediction, population reduction, trajectory integration and budget allocation) through seven replaceable modules that exchange conditional trajectory laws, whose award probabilities and award-marked moments give payment and uplift for any activity rule. A response-correction step reweights trajectories from abundant no-offer history to match the moments of a short pilot. We prove that, on a fixed plan menu and given the stage errors, the end-to-end value loss is bounded by the sum of four stage terms, and that for every stage there are instances on which omitting it leaves an error floor the others cannot remove. On 3,000 riders over 45 weekly origins, all 127 windows of a week are answered 11.04x faster with identical scenarios and at most 0.92% value lost by the allocation. On 24 new controlled response laws, the response correction with a one-week pilot lowers regret by 51.2% relative to a trial with the same nominal randomized rider-weeks, and a four-week pilot with exact summation comes within +0.007 of an 18-week trial. In registered studies where windows, populations, rules and binding budgets change from request to request, the framework's regret is below that of a trial with the same nominal rider-weeks and below dose interpolation of the same pilot data, and reusing its one-off preparation answers 60 requests 14.1x and 2.70x faster with identical answers. Against a nine-offer trial fitted with the framework's own dose curve, one-week regret is 0.055 lower.