Author: Zen and the Art of Computer Programming
1 Introduction
In recent years, with the development of emerging technologies such as mobile communication networks, the Internet of Things, and big data, the connection between service supply and demand has become closer and closer, so the demand response time in the logistics delivery process has become longer. Although early freight services were usually carried out by fixed appointment, suspension or delivery, with the development of economic scale and digitalization, the most popular now is the dynamic dispatch method based on the Internet, including map services, ride-hailing services, ride-hailing services and Shared bicycles, etc. These new supply and demand information exchange models bring huge opportunities to service providers and consumers, but also bring new challenges - how to accurately and timely disseminate demand information to users. How to utilize massive amounts of user demand information for accurate and efficient service scheduling is still an important issue.
For current scheduling problems, algorithmic methods based on deep learning have always occupied a research hotspot. Its biggest advantage is that it can automatically extract useful features from a large amount of data and learn its inherent patterns, so it can achieve efficient business decisions. This article will use deep learning algorithms to analyze and mine demand information from both service supply and demand parties, and then establish an accurate and fast service scheduling model. Specifically, the author designed a deep neural network model that comprehensively analyzes historical orders, user preferences and geographical location information, performs feature learning through abstract representation of user needs, and finally vector space mapping of features. Generate a transfer matrix of target user needs to effectively schedule users' actual needs. After the model training is completed, it can be applied to actual application scenarios to improve user experience and service quality.
The author hopes that through this paper, it can establish a more scientific and pragmatic thinking concept for scientific researchers in related fields and various companies in the industry, and use computer vision, machine learning, artificial intelligence and other scientific and technological means to provide solutions to the current supply and demand sides. lay a solid foundation for the information exchange market.
2.Related work
At present, there are several typical classification methods for scheduling models regarding supply and demand information exchange modes, such as rule-based model (Rule-based Model), predictive model (Predicti