Rebalancing in Bike-Sharing Systems – Incentives and Other Stories

on YouTube ↗43m

" summary

Station performance is modeled with Poisson flows and each station as a continuous-time Markov chain so the cost is the expected number of stock-outs over the day; that yields per-station 6 a.m. targets and an optimization for fleet placement. Moving about 500 docks (~2.5% of system) cut daily rebalancing costs roughly equal to that investment. A gamified bike angels program (10% invited, ~1% sign-up) awards points and one-week subscription extensions; crowd-sourced rebalancing matches motorized trucks and the live 30-minute clustering policy is what the system deploys.

AI-written summary · about

" chapters (7)

1. 0:00Start
2. 5:44System Imbalance
3. 11:36Planning Through Data
4. 14:22User Dissatisfaction Function
5. 15:34Neighborhood Effects
6. 38:52"REAL-TIME" INCENTIVES
7. 39:19SOME CONCLUDING THOUGHTS

" related talks

• 01h 08mUnderstanding and Shaping Urban Mobility Using Dollars and Data IBalaji Prabhakar · Real-Time Decision Making Boot Camp, Spring 2018
• 33mThe Rideshare Dispatch Problem (Aka Online Bipartite Matching with Supply Externalities)Sid Banerjee · Societal Networks 2018
• 41mDynamic Pricing in Ride-Sharing PlatformsRamesh Johari · Algorithmic Game Theory and Practice 2015
• 01h 38mRidesharingSid Banerjee · Real-Time Decision Making Boot Camp, Spring 2018
• 01h 03mUnderstanding and Shaping Urban Mobility Using Dollars and Data IIBalaji Prabhakar · Real-Time Decision Making Boot Camp, Spring 2018
• 53mCarpooling and the Economics of Self-Driving CarsMichael Ostrovsky · Matching and Objectives
• 08mTransforming Supply Chain: Smart LogisticsVenkata Anil Kumar Nilisetty · Conf42 Machine Learning 2025
• 52mDashboard Mechanisms for Online MarketplacesJason Hartline · Information Design and Data Science 2019