https://orcid.org/0000-0002-6396-0610
The central theme of this special issue is on Analysis, Design, and Optimisation in Smart and Connected Production and Service Systems. The smartness enables AI and advanced data analytics based intelligent operations and management in production and service, while the complexity is embedded through inter-connected components and units in the system. The purpose is to show the state-of-the-art research and applications in the general area of design and optimisation in smart and interconnected production and service, by bringing together researchers and practitioners from both academia and industry, to address the significant advancement, expose the unsolved challenges, present the needs for integration with new technologies, and provide visions for future research and development.
Due to strong competition in modern manufacturing and service, smart production and service become more critical, which is the foundation for sustainable success and competitiveness in the connected world. Substantial efforts have been devoted to research and practice of modelling, analysis, design, control, and optimisation. In recent years, significant advancement in information technology, particularly AI, machine learning, big data, etc., and the fast-growing economic activities, as well as the rapidly changing market have generated numerous opportunities for continuous improvement innovations. At the same time, many new challenges have emerged in order to apply and implement these innovations. Such opportunities and challenges have substantially expanded the scope of production research. In addition to manufacturing, the service sectors, such as healthcare, retail, energy, and transportation, are also experiencing paradigm changes. The first principle-based scientific methodology and rigorous quantitative approach are needed to analyse, design and optimise system structure and operations in such systems.
This special issue aims to publish original, significant and visionary papers describing scientific methods and technologies in analysis, design, and optimisation of smart and connected production and service systems with both solid the- oretical development, rigorous quantitative methods and models, and real world practice importance. A total of 66 submissions were received. Among them, 20 papers are finally included in this issue after a rigorous reviewing process. The contributions in this special issue can be divided into the following categories:
Performance analysis and optimization of production lines, job shop scheduling, supply chain and system design and decision, healthcare and logistics, and reviews.
The second category is on job shop scheduling, which has been one of the central topics in manufacturing system research. In paper ‘A novel feature selection for evolving compact dispatching rules using genetic programming for dynamic job shop scheduling’, Salama, Kaihara, Fujii and Kokuryo propose a new representation of Genetic Programming rules that abstract the importance of each terminal, and develop an adaptive feature selection mechanism to estimate terminals’ weights from earlier generations in restricting the search space of a current generation. It is demonstrated that the proposed approach outperforms other methods from the literature in a shorter computational time without sacrificing solution quality. In article ‘Deep reinforcement learning for dynamic scheduling of a flexible job shop’, Liu, Piplani and Toro present a hierarchical and distributed architecture to solve the dynamic flexible job shop scheduling problem. Using double deep Q-network algorithm to train scheduling agents, the relationship between production information and scheduling objectives is captured and real-time scheduling decisions can be made. In paper ‘An exact algorithm for an identical parallel additive machine scheduling problem with 2 multiple processing alternatives’, Kim and Kim introduce a bi-level optimisation model whose upper level problem is to determine a proper processing alternative for each product, and lower level to assign parts being produced to the additive machines. The exact algorithm consists of linear programming relaxation of a one-dimensional cutting stock problem, a branch-and-price algorithm, and a rescheduling algorithm, to find the optimal solution. Furthermore, in paper ‘Scheduling of autonomous mobile robots with conflict-free routes utilising contextual-bandit-based local search’, Jun, Choi and Lee divide the autonomous mobile robots (AMRs) problem into three sub-problems: path finding, vehicle routing, and conflict resolution. A comprehensive framework is presented to minimise total tardiness of transportation requests with consideration of conflicts between routes. Based on the shortest paths for vehicle routing, a new local search algorithm utilising the contextual bandit is proposed to select the best operator in consideration of contexts. Then an agent-based model with states and protocols resolves collisions and deadlocks in a decentralised way.
The third category covers supply chain and related system design and decisions. In paper ‘Dynamic inventory replenishment strategy for aerospace manufacturing supply chain: Combining reinforcement learning and multi-agent simulation’, Wang, Tao, Peng, Brintrup, Kosasih, Lu, Tang and Hu develop a multi-agent simulation model combined with a reinforcement learning-based dynamic inventory replenishment strategy to maximise the supply chain performance.
The approach has been applied in an aerospace manufacturing case study, which empirically demonstrates that the dynamic strategy yields considerable improvements and has an additional benefit of adaptivity to changes, such as demand and supply uncertainties. In paper ‘A regeneration process chain with an integrated decision support system for individual regeneration processes based on a virtual twin’, Kellenbrink, Nubel, Schnabel, Gilge, Seume, Denkena and Helber presents a multi-disciplinary case study of design and operation of a cyber-physical system demonstrator for individual, flexible and economically optimised maintenance, repair, and overhaul actions on extremely valuable components of aircraft engines. It operates with a virtual layer with a project-scheduling-based decision support system and uses a virtual twin of the object to be regenerated. In article ‘Big data-enabled intelligent synchronization for the complex production logistics system under the opti-state control strategy’, Zhang, Qu, Zhang, Zhong and Huang use simulation data of system operation to mine the relationship between the uncertain factors impact degree (UFID) and the system states, then use wrapper GA-DNN (Deep Neural Network) feature selection and classification method evaluates the UFID, which will be applied to the synchronisation decision. The results show that the method can accurately evaluate the UFID and avoid waste of resources and increase in operating costs, thereby also improves effectiveness and efficiency of opti-state control strategy. In article ‘A textual data-driven method to identify and prioritize user preferences based on regret/rejoicing perception for smart and connected products’, Du, Liu and Duan dig customer needs and evaluations from online customer reviews, then design a new directional distance index-based approach to acquire user weights. Combining absolute and three relative weights, we introduce an integrated approach to prioritise all customer preferences. An application of 12 kinds of smart speakers is constructed and discussed to illustrate the feasibility and usefulness of the proposed approach, and these corresponding results are helpful for smart design, development and improvement of smart, connected products.
The next category is related to service industry, such as healthcare and logistics.
In paper ‘A multidisciplinary approach to the development of digital twin models of critical care delivery in intensive care units’, Zhong, Sarijaloo, Prakash, Park, Huang, Barwise, Herasevich, Gajic, Pickering and Dong propose a qualitative and quantitative coupling approach to developing a digital twin model to investigate critical care delivery in intensive care units (ICUs). Such a model can be used to simulate real-life events as a full-fledged digital twin of the system, and as an in-silico testbed to investigate the real-time allocation of ICU resources such as medical equipment, flexible staffing, workflow change, and support decisions of patient admission, discharge, and transfer, for healthcare delivery innovation. In article ‘Managing appointments of outpatients considering the presence of emergency patients: the combination of the analytical and data-driven approach’, Wang and Liu formulate the patient appointment scheduling problem with the presence of emergency patients as a stochastic programming model to reduce the patient waiting time and increase server utilisation. Two methods are proposed to evaluate the patients waiting times and server utilisation for a given appointment schedule, and a simulated annealing algorithm is used to solve the SP model.
Moreover, in paper “A distributionally robust optimisation for COVID-19 testing facility territory design and capacity planning,” Fan and Xie develop a decision support tool for city governments to determine districts for testing facilities and their capacities based on the stochastic testing demand during a disease outbreak, by using a set-partitioning model embedded with a two-stage distributionally robust optimisation method.
In addition to healthcare, e-commerce has become an important element in the modern economy. To minimise the makespan under thread constraints and order precedence constraints in multi-thread fulfilment process of electronic order, in paper ‘Order processing task allocation and scheduling for e-order fulfillment’, Chen, Kang, Kang, Qi and Hu formulate the task allocation and scheduling problem as a mixed integer programming model and propose a novel depth-first heuristic to solve it. Extensions on precedence constraint reduction and resource allocation are also discussed to further improve and manage the e-order fulfilment process. Moreover, in article ‘Electric vehicle routing problem with flexible deliveries’, Sadati, Bkbari and Catay introduce flexible deliveries model where a fleet of EVs are used to serve customers and recharge their batteries along their routes. A hybrid Variable Neighbourhood Search coupled with Tabu Search are used to solve the problem of minimising total travel distance with minimum number of EVs. A case study in Nottingham, UK, is presented to provide insights. Furthermore, in paper ‘Smart logistics ecological cooperation with data sharing and platform empowerment: An examination with evolutionary game model’, Liu, Long, Wei, Xie, Wang and Liu explore the trend of ecological cooperation between the logistics platform and the supplier from the perspective of business ecosystem. Evolutionary game theory is used to describe the multi-period game between the two parties, and the equilibrium where the two parties both choose ecological cooperation and its preconditions is obtained. Specific paths to convert other evolutionary stable strategies into ideal one can be found by adjusting parameters, and suggestions for the platform to induce 4 its supplier to choose ecological cooperation can be provided.
The last category is on reviews. In paper ‘Deep reinforcement learning in production systems: a systematic literature review’, Panzer and Bender provide an overview of applications to motivate implementations and research of deep reinforcement learning (RL) supported production systems. It reveals that conventional methods are outperformed and implementation efforts on human experience are reduced, and future research could focus on analyzing safety aspects and demonstrate reliability under prevailing conditions. In the article ‘Distributing decision-making authority in manufacturing – review and roadmap for the factory of the future’, Antons and Arlinghaus study the concept of autonomy in production planning and control, enabled by cyber-physical systems and the distribution of decision-making authority. The perception of autonomy, technological requirements and increasing complexities of modern smart manufacturing are analyzed.
The advantages and benefits of autonomous control concepts are suggested and a distinct lack of quantitative results is identified. The Guest Editors would like to thank all the authors and reviewers for their outstanding work. We would also like to thank Editors-in-Chief Alexandre Dolgui, and many others for their efforts devoted to this Special Issue.