Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Review of Cloud Service Composition for Intelligent Manufacturing

Cuixia Li, Liqiang Liu, Li Shi

TL;DR

The paper addresses cloud-service optimization for intelligent manufacturing (IMfg-CSCOS) by defining 11 standardized optimization indicators across service demanders, providers, and platforms, and by consolidating optimization methods into heuristic and reinforcement-learning (RL) families. It surveys a wide range of algorithms, including NSGA-II, GA, PSO, ACO, SA, and advanced RL methods like DDPG and DQN, and discusses evaluation metrics such as hypervolume and convergence. A key contribution is the formalization of common objectives, constraints, and notations to improve reproducibility and comparability, along with a synthesis of research hotspots (service collaboration, dynamic environments, RL integration) and practical application scenarios in machining, automotive, and aerospace. The paper further identifies gaps—data availability, standardized simulators, and theoretical analysis—and outlines directions for future work to advance sustainable IMfg-CSCOS implementations with three-party stakeholder alignment and data-driven decision-making.

Abstract

Intelligent manufacturing is a new model that uses advanced technologies such as the Internet of Things, big data, and artificial intelligence to improve the efficiency and quality of manufacturing production. As an important support to promote the transformation and upgrading of the manufacturing industry, cloud service optimization has received the attention of researchers. In recent years, remarkable research results have been achieved in this field. For the sustainability of intelligent manufacturing platforms, in this paper we summarize the process of cloud service optimization for intelligent manufacturing. Further, to address the problems of dispersed optimization indicators and nonuniform/unstandardized definitions in the existing research, 11 optimization indicators that take into account three-party participant subjects are defined from the urgent requirements of the sustainable development of intelligent manufacturing platforms. Next, service optimization algorithms are classified into two categories, heuristic and reinforcement learning. After comparing the two categories, the current key techniques of service optimization are targeted. Finally, research hotspots and future research trends of service optimization are summarized.

Review of Cloud Service Composition for Intelligent Manufacturing

TL;DR

The paper addresses cloud-service optimization for intelligent manufacturing (IMfg-CSCOS) by defining 11 standardized optimization indicators across service demanders, providers, and platforms, and by consolidating optimization methods into heuristic and reinforcement-learning (RL) families. It surveys a wide range of algorithms, including NSGA-II, GA, PSO, ACO, SA, and advanced RL methods like DDPG and DQN, and discusses evaluation metrics such as hypervolume and convergence. A key contribution is the formalization of common objectives, constraints, and notations to improve reproducibility and comparability, along with a synthesis of research hotspots (service collaboration, dynamic environments, RL integration) and practical application scenarios in machining, automotive, and aerospace. The paper further identifies gaps—data availability, standardized simulators, and theoretical analysis—and outlines directions for future work to advance sustainable IMfg-CSCOS implementations with three-party stakeholder alignment and data-driven decision-making.

Abstract

Intelligent manufacturing is a new model that uses advanced technologies such as the Internet of Things, big data, and artificial intelligence to improve the efficiency and quality of manufacturing production. As an important support to promote the transformation and upgrading of the manufacturing industry, cloud service optimization has received the attention of researchers. In recent years, remarkable research results have been achieved in this field. For the sustainability of intelligent manufacturing platforms, in this paper we summarize the process of cloud service optimization for intelligent manufacturing. Further, to address the problems of dispersed optimization indicators and nonuniform/unstandardized definitions in the existing research, 11 optimization indicators that take into account three-party participant subjects are defined from the urgent requirements of the sustainable development of intelligent manufacturing platforms. Next, service optimization algorithms are classified into two categories, heuristic and reinforcement learning. After comparing the two categories, the current key techniques of service optimization are targeted. Finally, research hotspots and future research trends of service optimization are summarized.
Paper Structure (11 sections, 29 equations, 5 figures, 2 tables, 3 algorithms)

This paper contains 11 sections, 29 equations, 5 figures, 2 tables, 3 algorithms.

Table of Contents

  1. Introduction
  2. IMfg-CSCOS Process
  3. IMfg-CSCOS Objectives
  4. Service Optimization Methods
  5. Heuristic Optimization Algorithms
  6. RL-based Optimization Algorithms
  7. Additional Optimization Algorithms
  8. Research Hot Topics
  9. Application Scenarios
  10. Summary
  11. Declarations

Figures (5)

  • Figure 1: IMfg-CSCOS flow chart
  • Figure 2: Heuristic algorithm flow chart
  • Figure 3: NSGA-II algorithm flow chart
  • Figure 4: RL flow
  • Figure 5: DQN algorithm flow chart