A method and device for crowd-sourcing task allocation based on deep reinforcement learning

Abstract

The application provides a crowdsourcing task allocation method and device based on deep reinforcement learning, and aims at a crowdsourcing task allocation scene containing multiple heterogeneous tasks and multiple heterogeneous task undertakers, constructs an optimization function with the goal of solving a task allocation matrix on the basis of maximizing net profits of a task publisher and satisfaction degrees of the task undertakers, and then converts an optimization solving problem of the optimization function into a deep reinforcement learning problem of the task allocation matrix, so as to realize crowdsourcing task allocation. The application improves the solving speed of the task allocation matrix, and the solving result of the task allocation matrix is more suitable for reality.

Images