Research And Implementation On Taxi Cruise Path Recommendation Method Based On Pick-up Hotspots Areas

In recent years, with the rapid improvement of China’s urbanization level and the quality of life of urban residents, people put forward higher requirements for the efficiency of travel. With its flexibility and convenience, taxi becomes an important transport in people’s daily life. However, the main business model of domestic taxies called "Roadside Beckon" which is quite simple and extensive has great randomness. Passengers and the driver are ignorant of the location information of each other. Taxi drivers can only rely on their past experience to search for a cruise path. The status of "passive waiting" for taxies, causes the phenomenon of "free taxies cluster" and "a taxi is hard to find" in some places often occurs. This phenomenon has not only caused the increase of taxi empty rate and the decrease of taxi yields, but also led to the negative effects of urban traffic, deterioration of the traffic jam, environmental pollution and waste of resources etc.In recent years, some of the urban taxi-dispatching centers and taxi companies have adopted the artificial voice dispatching method, which is in a telephone reservation mode. It has effectively improved the ability of reasonable allocation of taxi resources. Meanwhile, the rise of various types of taxi Apps also have made an effort to solve the problem of "difficult to take a taxi". However, these methods have not fundamentally gotten rid of the taxi status of "passive waiting for passengers" Therefore, it urgently needs a strategy which is used to reasonably guide and dispatch taxis, so as to change the current situation of aimless random roaming for guest and balance the relationship between taxis and passengers.According to the above analysis, this dissertation proposes a taxi cruise path recommendation method based on pick-up hotspots areas. The method can recommend a cruise path with short distance and high pick-up rate for taxis. As a result, it solves the problem of "passive waiting" and aimless random roaming for guests in inefficient ways. The method is tested with real data in the dissertation and it provides a graphical result in the end. Meanwhile, the dissertation designs a taxi application based on the Android platform after implements the method. Specifically, the main work in this dissertation is as follows:Propose to make use of the improved grid clustering algorithm to cluster the taxi pick-up hotspots areas. This method, which has taken the historical pick-up spot data as the data source, clusters the taxi pick-up hotspots areas and finds the core of each area. It uses the Graham Scan algorithm and the Ray algorithm to calculate the pick-up rate of corresponding pick-up hotspots areas.In order to improve the efficiency of the cruise paths recommendation, this dissertation proposes to complete the process of the recommendation with two different models, one is the off-line path construction algorithm and the other is the online recommendation algorithm. This paper constructs all possible cruise paths with core pick-up spots, and calculates the expected cruise distance through Potential Travel Distance (PTD) equation. During the process of construction, it introduces two pruning methods to get rid of redundant paths. Results show that the recommendation process effectively improves the operation efficiency of the system.This dissertation collects the data of taxis in Android mobile intelligent terminals. It uses the technology of Global Positioning System (GPS), Geographic Information System (GIS), network communication, and the Message Queuing Telemetry Transport (MQTT) push to accquire the data, and deals these data with the operations of data preprocessing, storage and analysis. As a result, it provides truthful and accurate data source for the process of cruise paths recommendation.This dissertation realizes the cruise path recommendation method in program. Then, it designs and implements an independent call-response taxi system. The method is applied in the functional design of the driver terminal in this system, and tested on the Android mobile client. Finally, it achieved the good result.Based on the real taxi GPS informations, this dissertation randomly generates the position of the taxi which posts a cruise path remommendation request. Then, it programs the process of the cruise path recommendation based on pick-up hot spots and shows the final recommendation results in graphic in the end. This method provides great efforts to reduce the average empty loading rate of taxis, improves the operational efficiency of the taxi company and improves urban traffic conditions.