What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to discover and combine crowdsourced services moving with time and space efficiently and effectively (moving crowdsourced services), so as to ensure that users can continuously obtain a high - quality service experience during the movement. Specifically, the research focuses on the following aspects: 1. **Connectivity**: Mobile services need to maintain a connection with service providers, that is, they must be within the connection range of the providers. This requires the determination of co - movement patterns. 2. **Service Continuity**: Since service providers and users do not necessarily share all routes, an effective combination method is required to select the optimal set of available mobile crowdsourced services to ensure service continuity. 3. **Indexing Problem**: Existing co - movement discovery methods usually rely on centralized indexing methods, such as R - tree. As the scale of the data set expands, the performance of these methods will decline significantly. Therefore, a method without indexing is required for effective service discovery and combination. ### Research Challenges - **Connectivity**: Mobile services should maintain a connection with service providers, that is, they must be within the connection range of the providers. This requires the determination of co - movement patterns. To this end, the authors propose a service discovery algorithm based on parallel flocks, which uses spatio - temporal MapReduce to efficiently find flock patterns. - **Service Continuity**: When service providers and users do not fully share routes, how to ensure service continuity? An effective combination method is required to select the optimal set of available mobile crowdsourced services. - **Indexing Problem**: Existing co - movement discovery methods usually rely on centralized indexing methods, such as R - tree, but these methods perform poorly on large - scale data sets. Therefore, the authors propose a pruning method to avoid the selection of invalid services without building an index. ### Solutions To solve the above problems, the authors propose the following methods: 1. **A combination algorithm based on deep reinforcement learning**: It is used to discover and combine effective mobile crowdsourced services, taking into account quality - of - service (QoS) parameters. 2. **A parallel - flock - based service discovery algorithm**: As a baseline method, it is used to measure the accuracy of the proposed method. This algorithm is based on spatio - temporal MapReduce and can efficiently find flock patterns. 3. **An index - free discovery and combination framework**: It improves efficiency by using a pruning method to avoid the selection of invalid services. ### Experimental Verification Through experiments on two real - world data sets, the results show that this method is highly efficient and effective in terms of accuracy, learning speed, and scalability. ### Formula Examples The formulas involved in the paper are as follows: - **Capacity Calculation Formula**: \[ q_{\text{cap}}=\frac{B}{K}\log_2(1 + \text{str}) \] where: - \(B\) is the total available bandwidth, - \(K\) is the maximum number of concurrent requests that the mobile service MS can support, - \(\text{str}\) is the signal strength. - **Signal Strength Calculation Formula**: \[ \text{str}=\begin{cases} 1 & \text{if }0\leq p_{\text{dis}}(T_u.p_i, MS.p_i)\leq R_c\\ e^{-kd} & \text{if }p_{\text{dis}}(T_u.p_i, MS.p_i)>R_c \end{cases} \] where: - \(d = p_{\text{dis}}(T_u.p_i, MS.p_i)-R_c\), - \(p_{\text{dis}}(T_u.p_i, MS.p_i)\) is the vertical distance from the center point \(MS.p_i\) of the crowdsourced service to the user trajectory point \(T_u.p_i\), - \(R_c\) is the confidence radius, - \(k\) is the attenuation factor.

A Deep Reinforcement Learning Approach for Composing Moving IoT Services

Intelligent Orchestrating of IoT Microservices Based on Reinforcement Learning

Memory-based crowd-aware robot navigation using deep reinforcement learning

Human-Following and -guiding in Crowded Environments using Semantic Deep-Reinforcement-Learning for Mobile Service Robots

A Deep Learning Spatiotemporal Prediction Framework for Mobile Crowdsourced Services

Iraf: A Deep Reinforcement Learning Approach for Collaborative Mobile Edge Computing IoT Networks.

Collaborative multi-agents in dynamic industrial internet of things using deep reinforcement learning

Joint Service Migration and Resource Allocation in Edge IoT System Based on Deep Reinforcement Learning

Crowd-Aware Socially Compliant Robot Navigation via Deep Reinforcement Learning

Incentivizing Mobility of Multi-agent Vehicle Swarms with Deep Reinforcement Learning for Sensing Coverage Optimization.

Deep Reinforcement Learning for Social-Aware Edge Computing and Caching in Urban Informatics

A Reinforcement Learning Approach for Re-allocating Drone Swarm Services

Robot navigation in a crowd by integrating deep reinforcement learning and online planning

Crowd-Aware Mobile Robot Navigation Based on Improved Decentralized Structured RNN via Deep Reinforcement Learning

Deep Reinforcement Learning based Approach for Online Service Placement and Computation Resource Allocation in Edge Computing

Extracting microservices from monolithic systems using deep reinforcement learning

A Composite Service Selection Approach Based on Functionality Folding and Reinforcement Learning

Logistics-involved QoS-aware service composition in cloud manufacturing with deep reinforcement learning

A Deep Reinforcement Learning Approach for Security-Aware Service Acquisition in IoT

Mobility and energy efficient services composition algorithm with QoS guarantee for large scale Cyber–Physical–Social Systems

Proactive Composition of Mobile IoT Energy Services