Study On Joint Decision Of Location Of Distribution Point And Dynamic Pricing Of Shared Resource

In the new industry of sharing economy,common shared resources include shared bicycles,shared cars,shared umbrellas,shared charging pads,etc.Due to the rapid development of these shared resources in the field of shared industry,the research problems and research methods for the location of Shared resource distribution points at home and abroad are relatively simple,and the premise of solving the problems is to clarify the demand function of shared resources to a certain extent.In this paper,based on these limitations,considering the coupling between facility location and pricing,the problem to be studied becomes a compound problem.That is,the Reinforcement Learning method is used to solve the joint decision problem of distributed point location and dynamic pricing of shared service when the shared resource demand function is unknown.In this paper,based on the theoretical knowledge of facility location and dynamic pricing,the problems that need to be studied are proposed,and the methods of Reinforcement Learning are introduced.The reinforcement learning system and its agent learning process are elaborated in detail.At the same time,the steps of Reinforcement Learning algorithm are described and the experimental process is designed.The research objective of this paper is to make sharing resource operators get the maximum benefits.The research problems are mainly divided into two aspects:(1)joint decision research on location of single sharing resource distribution point and dynamic pricing of sharing service.(2)joint decision research on location of Shared resource compound distribution point and dynamic pricing of Shared service.In the case of unknown demand function and user behavior characteristics,the joint decision problem of location of Shared resource distribution point and dynamic pricing of Shared service is firstly elaborated.Then the Reinforcement Learning algorithm model is established according to the problem,and the state,behavior and reward function are defined.Then a specific Reinforcement Learning algorithm is established according to the goal of Reinforcement Learning decision.Finally,based on the Reinforcement Learning algorithm,the experimental process is designed,experimental parameters are set and experiments are run to verify the effect of reinforcement learning with examples.With the increasing number of iterations in the experiment,the learning effect of the agent is getting better and better,the learning curve increases asymptotically and the learning growth rate increases obviously,thus the enterprise income is improved.After repeatedly designing the experimental process and adjusting the experimental parameters,it is found that the learning curve of the agent in the continuous interaction with the environment to select the distribution point and the dynamic pricing joint decision-making has reached a asymptotically stationary state through the experimental iteration.In the joint decision-making problem of single distribution point location and dynamic pricing,the reward is increased by about 40%.In the joint decision-making problem of compound location and dynamic pricing,the reward growth rate is about 80%.The experimental results show that the reward growth rate of multi-distribution point compound location and dynamic pricing joint decision is higher than that of single distribution point location and dynamic pricing joint decision.At the same time,it also shows that enterprises will get greater corporate income in the choice of multiple distribution points.The experimental results show that the agent in the Reinforcement Learning method has the ability to capture and adapt to the change of the environment in the process of learning.It reflects the applicability of the Reinforcement Learning method to solve the problem of joint decision-making on the location of shared resource distribution points and the dynamic pricing of shared services,and also provides management decision suggestions for the operators of shared resources.