Order assignment method and system based on interactive reinforcement learning

Abstract

The invention relates to the technical field of internet information, in particular to an order assignment method and system based on interactive reinforcement learning. According to the invention, man-machine interaction is introduced in an autonomous learning process, a man-machine interaction mode of human demonstration, interference and evaluation is fused, learning is carried out from the human demonstration, and a real order assignment scene can be better simulated through real data of the human demonstration; learning is carried out from human interference, when wrong actions occur in the autonomous learning process, the performance of the intelligent agent is controlled, and wrong results are avoided; learning is carried out from human evaluation, an autonomous learning result is evaluated manually, so that the learning process shifts to a better order assignment strategy direction, the learning process is accelerated, and an optimal order assignment strategy is obtained.

Description

technical field [0001] The invention relates to the field of Internet information technology, in particular to an order assignment method and system based on interactive reinforcement learning. Background technique [0002] Online ride-hailing apps and platforms have become a new and popular way to provide on-demand transportation services through mobile apps. At present, some taxi-hailing mobile applications such as Didi, Uber, and Lyft are popular all over the world. The system provides services for a large number of passengers every day and generates a large number of taxi-hailing orders. For example, China's largest online car-hailing service provider Didi needs to process about 11 million orders every day. The order assignment problem of online taxi service is essentially a reasonable match between potential passengers and drivers. By. In many cases, the service is reusable, and the service provider will disappear for a period of time after being matched with the user...

AI Technical Summary

Problems solved by technology

[0004] The existing technologies basically carry out completely autonomous training and learning process through the interaction between traditional reinforcement learning and the environment. However, this completely autonomous learning method lacks human participation, and the learning process takes a lot of time; the learning process cannot be controlled. The behavior of the agent may produce wrong results; the learning results are difficult to simulate complex real scenes

Images