Mobile Opportunistic Planning: Methods and Models

MSR-TR-2006-184 |

Publication

We describe the construction of probabilistic user models that predict the context-sensitive cost of time, and the use of these models within a system that deliberates about adding waypoints to trips opportunistically so as to satisfy standing goals. We present details about the use of the models in a prototype named Mobile Commodities (MC). When travel is sensed via GPS, MC performs an ongoing analysis about the cost of diverging from a direct path to a destination. The system considers the location of potential waypoints that can satisfy goals, pricing of items at waypoints, and costs associated with the addi-tional time and travel required for the modified trips.