Abstract:Extensive research has been carried out on reinforcement learning methods. The core idea of reinforcement learning is to learn methods by means of trial and error, and it has been successfully applied to robotics, autonomous driving, gaming, healthcare, resource management, and other fields. However, when building reinforcement learning solutions at the edge, not only are there the challenges of data-hungry and insufficient computational resources but also there is the difficulty of a single reinforcement learning method to meet the requirements of the model in terms of efficiency, generalization, robustness, and so on. These solutions rely on expert knowledge for the design of edge-side integrated reinforcement learning methods, and they lack high-level system architecture design to support their wider generalization and application. Therefore, in this paper, instead of surveying reinforcement learning systems, we survey the most commonly used options for each part of the architecture from the point of view of integrated application. We present the characteristics of traditional reinforcement learning in several aspects and design a corresponding integration framework based on them. In this process, we show a complete primer on the design of reinforcement learning architectures while also demonstrating the flexibility of the various parts of the architecture to be adapted to the characteristics of different edge tasks. Overall, reinforcement learning has become an important tool in intelligent decision making, but it still faces many challenges in the practical application in edge computing. The aim of this paper is to provide researchers and practitioners with a new, integrated perspective to better understand and apply reinforcement learning in edge decision-making tasks.

THE METHOD OF SELF-REINFORCEMENT OF INDEPENDENT-EDGES

An Edge Extraction Technique Using the Accumulation of Self-Reinforcement of Search Trajectories of Independent-Edges

A Novel Method For Image Edge Extraction

Edge Segmentation Based on Automatic Scale Selection

A Reinforced Road Detection Method in Complicated Environment

Edge Extraction of Face Images for Expression Analysis

A Stochastic Heuristic Search Algorithm for Suppressing Noise in Edge Extraction

An Application Of Stochastic Heuristic Search Method To Edge Extraction In Noisy Images

SuperEdge: Towards a Generalization Model for Self-Supervised Edge Detection

STEdge: Self-training Edge Detection with Multi-layer Teaching and Regularization

Study on Image Edge Detection Base on Strain-Loose Method

Continuous Multiresolution Modeling Based on Interpolation Subdivision

Edge- and vertex-reinforced random walks with super-linear reinforcement on infinite graphs

Interacting Edge-Reinforced Random Walks

Automatic Image Edge Extraction and Segmentation by Random Seed Region Search Algorithm

Reinforcement-Tracking: An End-to-End Trajectory Tracking Method Based on Self-Attention Mechanism

Linearly edge-reinforced random walks

How to Design Reinforcement Learning Methods for the Edge: An Integrated Approach toward Intelligent Decision Making

RULE-BASED EDGE LINKING ALGORITHM FOR EXPRESSWAY DISTRESS DETECTION

Deep reinforcement learning-based sampling method for structural reliability assessment

Adaptive Balancing of Edge Extraction in LADAR-Referenced Navigation over Plain Area