Blockchain Based Storage And Matching Method For Package Delivery Via Ridesharing

Ridesharing has been widely accepted by the market with its low prices and convenient services.However,most of the current ridesharing methods are based on a centralized architecture,which means there exists a third-party company in charge of collecting and publishing ridesharing-related information.By this means,sharing participants need to upload sensitive information such as identity and location to the server.Meanwhile,third-party companies may modify the matching result to gaining more profit.To avoid the influence of third parties on the ridesharing system,this article focuses on building a distributed ridesharing system.However,due to the lack of unified control of third-party nodes,it is difficult to effectively store the ridesharing information in a distributed manner.In addition,the lack of third-party nodes also brings challenges to ridesharing matching.Sharing participants can only rely on their local information to perform the matching process,which will result in a decrease in the matching efficiency.Meanwhile,most of the current ridesharing software is developed based on a central architecture,and it is not suitable for the distributed ridesharing services.This paper conducts research on the above challenges,and the main research work is summarized as follows.To tackle the security threats of distributed storage in the package delivery ridesharing system,this thesis proposes a Blockchain secured ridesharing delivery scheme and precisely describes the entire process from request generation to blockchain confirmation.This method utilizes the immutability and distributed architecture of the Blockchain and can effectively prevent data tampering.However,such tamper resistance property comes at the cost of a long confirmation delay caused by the consensus process.To improve the Blockchain efficiency,a Hash-oriented Practical Byzantine Fault Tolerance(PBFT)based consensus algorithm is proposed,which can reduce the transaction confirmation delay from 10 minutes to 15 seconds.Besides,the Hash-oriented PBFT effectively avoids the double-spending attack and Sybil attack.Security analysis and simulation results demonstrate that the proposed Blockchain secured ridesharing delivery system offers strong security guarantees and satisfies the quality of delivery service in terms of confirmation delay and transaction throughput.To resolve the matching efficiency problem resulted from limited computing resources and request information,this thesis proposes an event-triggered distributed deep reinforcement learning(ETDDRL)algorithm.The proposed method can generate/update the real-time ridesharing orders for the new coming ridesharing requests from a local view.As the vehicle only has limited computing capability,this thesis first derives the approximate searching range of the vehicle under a certain detour rate,thereby reducing the computation consumption of the vehicle.Subsequently,this thesis focuses on abstracting the matching model,taking the picking up event of new packages as a condition,this thesis proposes a state transition method based on event triggers,which solves the mapping problem between multi-path matching and reinforcement learning models.Meanwhile,considering the sequence problem between newly accepted requests and undelivered requests,this thesis proposes a sequential ant colony algorithm-based path planning algorithm,which can assist the state transition targeting on minimizing the traveling length.Furthermore,this thesis analyzes the driving data of 12,509 taxis in Beijing in about one month to imitate package delivery requests,these requests can be used to evaluate the matching efficiency of the proposed method.Simulation results demonstrate the vast potential of the ETDDRL-based matching algorithm,when the maximizing package receiving number is certain,this algorithm can achieve a maximum increase of 86.6%in terms of average profit per unit distance.In addition,this thesis also implements the related software involved in the package delivery ridesharing system.Firstly,this thesis implements the ridesharing client software based on the Android platform.Specifically,the background of the interface is based on the Baidu map layer,and this thesis also realizes the self-setting of the start and end positions and can display the current driving path of the vehicle.Besides,after each new request has been picked up,the software can dynamically update the current route according to the planning results of the inner ant colony algorithm,which meets the driver's route requirements.Subsequently,this thesis constructs a blockchain-based distributed storage software.According to the hierarchical structure of the blockchain,this article first constructs the data link layer and realizes a series of blockchain structures including blockhead,block body,and Merkel tree.Second,the User Datagram Protocol is used to realize the interconnection and intercommunication between different nodes,which mainly includes functions such as block information sending,block information receiving,and block information verification.Subsequently,a variety of consensus algorithms such as proof-of-work and proof-of-stake were supported at the consensus layer,which expanded the application scenarios of this system.Practical application tests show that the software implemented in this article can meet the requirements of actual use.