About the Project
Another area of application for these methodologies is to address the challenge of deploying resources in a real time manner, often without complete information of how the environment will evolve in the near future. This kind of real time optimization is often encountered in service platforms such as didi and grab, where the decisions, like matching drivers and passengers, need to be made in quick succession, to satisfy a number of KPIs like good customers service experience and rewarding good driver etc. This project will bring the core technologies developed here into the online and multi-objective world. We have deployed a system in a ride sharing platform in china, and has demonstrated that good pareto solution can actually be obtained in an online fashion, provided the unknown environment is reasonably stable. The challenge here is to extend this theory to the more dynamic and time varying environment.
