New AI Framework Generates Realistic Daily Schedules 52% More Accurately
Dynamic programming and travel time simulation slash schedule errors by over half.
Deep Dive
A new framework uses dynamic programming and simulated travel times to generate realistic individual activity schedules. The method iteratively refines activity location allocation, reducing discrepancies between simulated and survey-reported travel times by 52.2% relative to the first iteration. The work addresses challenges in mobility modeling for applications such as infectious disease control, urban transportation planning, and policy design.
Key Points
- Dynamic programming reduces travel time discrepancies by 52.2%
- Framework uses simulated travel times to iteratively refine activity locations
- Enables privacy-preserving generation of realistic individual schedules for urban planning and epidemiology
Why It Matters
More realistic synthetic schedules mean better disease models and smarter city planning without invading privacy.