A Multi-Task Service Composition Method Considering Inter-Task Fairness in Cloud Manufacturing

Cloud manufacturing is an emerging novel business paradigm for the manufacturing industry. In cloud manufacturing, distributed manufacturing resources are encapsulated into services and aggregated in a cloud manufacturing platform. Through centralized service management, cloud manufacturing is capable of dealing with multiple requirement tasks simultaneously. The ability to deal with multiple tasks at the same time is an important characteristic that distinguishes cloud manufacturing from the previous networked manufacturing models such as manufacturing grid. When it comes to multiple tasks in cloud manufacturing, a critical issue is how to schedule massive services to complete them with shortest makespan, lowest cost, and highest quality, etc. In order to facilitate the research on this issue, we in this paper propose a model for multitask-oriented service composition and scheduling in cloud manufacturing, in which key factures of cloud manufacturing such as service orientation, involvement of logistics, and dynamical change of service availability are taken into account. New concepts such as service efficiency, enterprise capability, and task workload are introduced, and various types of times including service time, logistics time, and waiting time are analyzed in detail. Moreover, this model can be conveniently extended by incorporating new elements such as task constraints, task priority, and continuous task arrival. An example that motivates the current model is presented. Simulation experiments with different numbers of tasks are performed to demonstrate the feasibility of the model.

Cloud manufacturing is a new service-oriented manufacturing model that integrates distributed manufacturing resources to provide on-demand manufacturing services over the Internet. Service composition that builds larger-granularity, value-added services by combining a number of smaller-granularity services to satisfy consumers’ complex requirements is an important issue in cloud manufacturing. Meta-heuristic algorithms such as genetic algorithm, particle swarm optimization, and ant colony algorithm are frequently employed for addressing service composition issues in cloud manufacturing. However, these algorithms require complex design flows and lack adaptability to dynamic environment. Deep reinforcement learning provides an alternative approach for solving cloud manufacturing service composition issues. This chapter proposes a deep Q-network (DQN) based approach for service composition in cloud manufacturing, which is able to find optimal service composition solutions through repeated training and learning. Results of experiments that take into account changes of service scales and service unavailability reveal the scalability and robustness of the DQN algorithm-based service composition approach.

Cloud manufacturing as an emerging service-oriented manufacturing paradigm integrates and manages geographically distributed manufacturing resources such that complex and highly customized manufacturing tasks can be performed cooperatively. The service composition and optimal selection (SCOS) problem, in which manufacturing cloud services are optimally selected for performing subtasks, is one of the key issues for implementing a cloud manufacturing system. In this paper, we propose a new mixed-integer programming model for solving the SCOS problem with sequential composition structure. Unlike the majority of previous research on the problem, in the proposed model, the transportation between distributed resources and its effects on quality of services are considered. Although a wide variety of metaheuristics have been tailored for solving the SCOS problem, no consistent and comprehensive conclusion has been reached so far on the superiority of a specific algorithm. Therefore, for the first time, solution space landscape of the problem was analyzed through several statistical criteria which demonstrated that the landscape is rugged and local optima are clustered in a small region of the search space. Therefore, to find good solutions, a metaheuristic algorithm needs to perform both proper exploitation and exploration of the search space. According to the landscape analysis, the basic imperialist competitive algorithm (ICA) was hybridized with a local search (LS) algorithm resulting in the hybrid ICA (HICA). To examine the performance of the proposed HICA, an example of online motorcycle production in the USA as well as four randomly generated large-scale instances, were solved through the LS, ICA, and HICA. Computational results showed that transportation consideration is indispensable for obtaining more realistic solutions in cloud manufacturing. The results also revealed that the HICA outperformed the LS and basic ICA in terms of the value of cost objective function, the stability of solutions and convergence speed. Hence, not only statistically but also analytically, it was proved that algorithms incorporating both exploitation and exploration are able to solve the SCOS problem more efficiently.

Cloud manufacturing (CMfg) is a new type of service-oriented manufacturing paradigm, in which all kinds of manufacturing resources are encapsulated as manufacturing services (MSs) and can be invoked by customers on demand. Manufacturing service composition (MSC) is a key technology in CMfg for creating value-added services to complete complicated manufacturing tasks by aggregating qualified MSs together. However, current MSC approaches have some drawbacks and there still exist some issues remained to be solved: (1) large quantities of candidate services increase the complexity of service dynamic composition, which poses scalability concerns and on-demand efficient solutions; (2) the service domain features (e.g., service prior, correlation, and similarity) that have a strong influence on the efficiency of service composition are not considered adequately, which causes undesirable efficiency in practical service applications; and (3) dynamic characteristics of QoS (quality of service) values in an open network environment are not considered adequately. To effectively address such problems, this paper first proposes a context-aware artificial bee colony (caABC) algorithm based on the principle of ABC and service features in the cloud environment. Then the differential evolution-enhanced caABC, i.e., the so-called DE-caABC, is designed to increase the searching performance of ABC further. Additionally, dynamics of trust QoS is investigated with the introduction of time decay function. Finally, the feasibility and effectiveness of DE-caABC are validated through the experiments.

Cloud manufacturing (CMfg) is a relatively new manufacturing paradigm which promotes easier collaboration among geographically-dispersed manufacturers. Various manufacturing resources can be shared via cloud manufacturing platform as services. The efficacy of such resource sharing and thus collaboration in this service-oriented mode of manufacturing is highly dependent to the efficiency of underlying mechanisms used for processing these services. Activities such as service matching, service retrieval, service composition and optimal selection, and service scheduling are among the most critical of those service processing mechanisms. In this paper, we introduce and implement a comprehensive service composition and optimal selection (SCOS) technique that takes advantage of a real-world manufacturing capability dataset, deep learning models as well as Word2Vec technique to retrieve appropriate candidate sets for each submitted manufacturing subtask in the cloud manufacturing platform. Then, a genetic algorithm is implemented to obtain a near-optimal composite service that achieves the highest Quality of Service (QoS). A series of experiments were conducted to prove the feasibility and efficacy of the approach and results were presented.

Chapter 1. Introduction to scheduling in Industry 4.0 and cloud manufacturing systems.- Chapter 2. Proactive scheduling and reactive real time control in Industry 4.0 .- Chapter 3. Using a digital-twin for production planning and control in Industry 4.0.- Chapter 4. Adaptive Scheduling in the era of Cloud Manufacturing.- Chapter 5. Cloud Material Handling Systems: conceptual model and cloud-based scheduling of handling activities.- Chapter 6. Coupling robust optimization and Model-Checking techniques for robust scheduling in the context of Industry 4.0.- Chapter 7. Integrated scheduling of information services and logistics flows in the omnichannel system.- Chapter 8. Human-oriented assembly line balancing and sequencing model in the Industry 4.0 era.- Chapter 9. A generic decision support tool to planning and assignment problems: Industrial applications & Industry 4.0.- Chapter 10. The Manufacturing Planning and Control system: a journey towards the new perspectives in Industry 4.0 architectures.- Chapter 11. Multi-criteria single batch machine scheduling under time-of-use tariffs.- Chapter 12. Service composition in cloud manufacturing: a DQN-based approach.- Chapter 13. The Tolerance Scheduling Problem in a single machine case.

Aim and ApproachLarge cooling water system has great potential of energy saving due to improper operation control. The coupled hydraulic and thermodynamic characteristics among chillers, pumps, and cooling towers lead to the difficulty of cooling water system operation. This study investigates an advanced control approach of multi-agent reinforcement learning (RL) for a large cooling water system. The multiple RL agents interact with each other to deal with the complex coupled characteristics in the cooling water system. The proposed approach is implemented in a real-world cooling water system and validated with experiments.Scientific Innovation and RelevanceIn this study, a multi-agent RL approach is proposed to separately control the chillers, pumps, and cooling towers in the cooling water system. The soft actor critic algorithm is used in the RL agent, as it is an efficient and advanced RL algorithm for continuous and discrete control problems. The agents are designed to have private states, private actions, and shared rewards. The reward functions are carefully designed to balance the agents for global optima.To provide a realistic environment for RL training, a detailed physical modeling platform of a cooling water system is established and validated. This platform is capable of hydraulic and thermodynamic calculation of cooling water systems with customized numbers and characteristics of chillers, pumps, and cooling towers, which is scalable for different cooling water systems.The proposed multi-agent RL controller is implemented in the building automation system of a real cooling water system. A one-month experimental study of the multi-agent RL performance is conducted. The proposed multi-agent RL controller is compared with a single-agent RL controller and a rule-based controller through real-world experiments to demonstrate the energy performance. This is a pilot study of multi-agent RL approach to real-world cooling water system control.Preliminary Results and ConclusionsThe preliminary simulation study is conducted in the established cooling water system model. The electricity consumption simulation error is 4.0% in coefficient of variation of mean absolute error (CVMAE) compared with the measured 15-minute electricity consumption. The 6-month simulation shows that the single-agent RL can save 6.2% energy from the rule-based control, while the multi-agent RL can save 7.2% electricity consumption of the cooling water system. In the following study, the multi-agent RL implementation will be practiced with more experiments and applications for cooling water system control.

Diffusion mechanism simulation of cloud manufacturing complex network based on cooperative game theory

Service composition and scheduling in cloud manufacturing (CMfg) has attracted increasing attention due to CMfg's platform's superior ability in integrating distributed resources and distributing integrated resources. Existing studies assume that the manufacturing cloud services (MCSs) of enterprises are all available throughout the planning horizon, yet an MCS may be occupied during specific time intervals. To deal with this issue, we study the multitask‐oriented service composition and scheduling problem (SCSP) in CMfg with unavailable time windows and subtask rejection (MTO‐SCSP‐UTWSR), whereby some subtasks of the CMfg platform's customers may be rejected, which will incur a penalty cost due to the limited service capacity of the MCSs. The customers are divided into subtask‐rejection‐sensitive customers and subtask‐rejection‐insensitive customers, and the total subtask rejection cost of each subtask‐rejection‐sensitive customer cannot exceed a predefined threshold. The objective is to determine the customer‐enterprise allocation scheme, the MCS composition for serving the subtasks allocated to each enterprise, and the subtask schedule for each MCS, subject to some service flexibility constraints, so as to minimize the sum of the customer‐enterprise allocation cost, opening cost of MCSs, and total subtask rejection cost of the subtask‐rejection‐insensitive customers. To solve the problem, we develop a Benders decomposition‐based hybrid algorithm in which the Benders subproblems are solved by a tailored branch‐and‐price‐and‐cut algorithm, which incorporates several enhancement techniques. Extensive numerical studies demonstrate the superiority of the developed algorithm in solving the problem and quantify the impacts of some model parameters on the solution structure and algorithm performance.

In the cloud manufacturing systems, both manufacturing tasks and manufacturing services are in a dynamic environment. How could cloud manufacturing platform optimizes manufacturing cloud services based on QoS, matching an optimal service composition for manufacturing tasks has become an urgent problem at present. In view of this problem, we study the matching of manufacturing tasks and manufacturing services from the perspective of complex network theory. On the basis of manufacturing task network and manufacturing service network, a dynamic matching network theory model of manufacturing task-service is constructed. And then, we take a dynamic assessment of QoS. Finally, we use load and dynamic QoS as the optimization objectivities, transform the optimal manufacturing service composition problem into the shortest path problem, and the dynamic scheduling of manufacturing services is realized.

A Multi-Agent Reinforcement Learning Framework for Public Health Decision Analysis.

In the cloud manufacturing environment, the sources of manufacturing tasks and resources are more extensive. The balanced utilization of manufacturing resources is conducive to the timely completion of tasks and the good operation of the cloud system platform. Based on the network graph theory, this paper firstly constructed the cloud manufacturing process network graph based on product manufacturing BOM, described the selection constraint relationship between equipment in each process of the product, and formed the matching relationship between product processing process and manufacturing equipment. When new cloud manufacturing tasks and cloud manufacturing resources are added, or the processed and pending tasks and cloud manufacturing resources change, the cloud manufacturing dynamic matching network is updated synchronously. Secondly, considering the load of manufacturing resources, a task load queue centered on manufacturing resources was constructed. The nonlinear programming model was used to build a workshop equipment scheduling model aiming at minimizing the total processing time and cost of products. Finally, genetic algorithm is used to solve and verify the validity and accuracy of the model.

Cloud manufacturing as a new service-oriented business model transforms manufacturing resources into manufacturing services and manages them centrally via Cloud manufacturing platforms. Using distributed manufacturing resources is an opportunity to decrease the logistics and production costs by allocating tasks to manufacturing resources toward task destination nodes. Decomposing ordered tasks to several subtasks, allocating distributed manufacturing resources to tasks, scheduling manufacturing tasks, and select appropriate logistics strategies to deliver tasks to service providers are important activities that must be considered to improve the performance of a Cloud manufacturing system. This paper proposes a mathematical model to improve the performance of the Cloud manufacturing systems. To that end, task decomposition is implemented based on the alternative process routing concept. Decomposing each task to several manufacturing plans results in enhancing system flexibility and consequently improve the performance of cloud manufacturing. Besides, a consolidated production-distribution toward destination nodes strategy is employed to allocate and deliver tasks to manufacturing resources at minimum costs. The impacts of the alternative process routing concept and consolidated production-distribution toward destination nodes strategy are discussed by considering different scenarios and the capabilities of the proposed model have been demonstrated.

In the cloud manufacturing platform, virtualized manufacturing resources are complex to ensure the authenticity and security of transaction data during the transaction process and rely on the unified deployment of cloud manufacturing platform operators (CPOs). This centralized framework is prone to a single point of failure and risks leakage of private data. Because of these problems, combined with the advantages of decentralization and immutability of blockchain, we innovatively explore the application of blockchain technology in cloud manufacturing and propose a blockchain-based cloud manufacturing resource allocation method. Firstly, a blockchain-based decentralized cloud manufacturing trading platform framework is presented. The Elliptic Curve Digital Signature Algorithm (ECDSA) in manufacturing resource/demand release is studied, and the matching mechanism between manufacturing resources and demand is analyzed. Then, a manufacturing resource verification contract and a manufacturing resource transaction contract for cloud manufacturing are designed using smart contracts (SC). The experimental tests in the Remix platform are completed using Solidity. The test results show that the proposed method can provide a safe and reliable guarantee for virtualized manufacturing resource transactions. Finally, the game problem of supply-demand balance between manufacturing resource suppliers (MRS) and demanders (MRDs) under the decentralized architecture is explored, and simulation is carried out on MATLAB R2021b. The simulation results show the manufacturing resource supply and demand. The fair game can reach the Nash equilibrium and has a faster convergence rate than the existing research.

Smart transportation networks are challenged by a critical challenge of traffic congestion and inevitably requires a complex package of management solutions. This paper introduces a Secure Multi-Agent Reinforcement Learning (MARL) framework that optimizes traffic whilst protecting from malicious threats. Traffic Prediction Dataset accessible through Kaggle is used as the foundation for training reinforcement learning agents from four junctions who cooperate to manage traffic signals. The widespread preprocessing procedures that also dealt with data gaps in addition to dimension transformation and value normalization generated genuine traffic prediction models. The Multi-Agent Reinforcement Learning framework reduced the mean waiting time by 45.7% with a 47.0% traffic passing rate higher than that achieved using standard fixed-duration signal control techniques. Homomorphic encryption, when integrated with blockchain protocols, instituted novel security protocols for agent to agent communication, and this yielded a reduction of adversarial attacks below 90% levels. The implementation framework provided a 39.3% reduction in the fuel consumption when implemented in real-time urban traffic management systems. Secure Multi-Agent Reinforcement Learning (MARL) systems show their potential aptitude to propel smart transportation networks in the while providing increased efficiency and security as well as sustainability effectiveness.