Workload-based multi-task scheduling in cloud manufacturing
Introduction
Cloud manufacturing is a new service-oriented business model aiming for sharing and collaboration of large-scale manufacturing resources [1], [2]. It realizes its objective through establishment of a common cloud manufacturing platform, which aggregates distributed manufacturing resources encompassed in the entire product life cycle, transforms them into manufacturing services, and manages them centrally [3], [4]. Through centralized management and operation of services, cloud manufacturing is able to deal with multiple requirement tasks at the same time. A critical issue with cloud manufacturing is therefore how to schedule multiple tasks to achieve optimal system performance. Different from the scenario in cloud computing, task scheduling in cloud manufacturing is usually accompanied by logistics. The involvement of logistics makes the multi-task scheduling in cloud manufacturing more complicated.
Multi-task scheduling in cloud manufacturing refers to process of allocating services over time to perform a set of tasks while satisfying constraints in terms of time, cost, QoS, and service availability. Task scheduling is an intrinsic part of a cloud manufacturing system, and has a major impact on system performance. Effective task scheduling methods are capable of significantly enhancing system performance. For multi-task scheduling, scheduling objective should be achieving the overall optimization of all tasks. Multi-task scheduling requires the consideration of coupling relationships (e.g. different tasks may require the same type of services) among multiple tasks. Traditional methods for single-task scheduling may not achieve the optimal system performance under multi-task scenarios as they do not deal with all task as a whole [5], [6], [7], [8]. Multi-task scheduling in cloud manufacturing has been considered in literature [9], [10], [11]. However, these works dealt with either only homogeneous tasks or using a different model and method. Multi-task scheduling in cloud computing has also been studied [12], [13]. However, due to the fundamental differences between cloud manufacturing and cloud computing [4], [14], the proposed approaches cannot be applied directly to cloud manufacturing. It is therefore necessary to explore new, effective methods for multi-task scheduling in cloud manufacturing.
In this paper, we address the issue of multi-task scheduling in cloud manufacturing based on task workload [13]. The innovations of this work are as follows. First of all, we proposed a new multi-task scheduling model for cloud manufacturing based on service composition idea and method. Some critical issues pertaining to scheduling in cloud manufacturing such as logistics are taken into account. Secondly, the proposed model incorporates novel methods for modelling task workload and service (including service quantity and efficiency) [15], [16], which enables us to dynamically calculate task (or subtask) fulfilment time and service utilization [17]. More importantly, based on different workload-based task scheduling methods, we find that scheduling larger-workload tasks with a higher priority can lead to better system performance such as a shorter makespan and higher service utilization. Monte Carlo methods are employed to reveal the regularity behind the scheduling methods [9], [10], [11].
The rest of this paper is structured as follows. In Section 2, a systematic literature review and corresponding analysis are conducted. Section 3 gives an example that motivates the establishment of the current multi-task scheduling model. Section 4 elaborates on the multi-task scheduling model in detail. In Section 5, a concrete multi-task scheduling example is given. Section 6 presents the results of simulation experiments and associated analysis. And finally, Section 7 concludes this paper followed by discussions on future research.
Section snippets
Literature review
First of all, it is necessary to clarify a number of fundamental concepts such as manufacturing tasks, resources, and services. Wang et al. [18] discussed the manufacturing task semantic modelling and description in a manufacturing system. In their view, manufacturing tasks can be divided into nine categories, including design tasks, manufacturing and processing tasks, logistics and inventory tasks, etc. A manufacturing task information model consisting of static information, subtask set,
A motivating example
Consider a cloud manufacturing platform involving 10 enterprises from Guangdong Zhaoqing Automotive Parts Industry Association, which are Huaiji Dengyun Auto-parts (Holding) Co., Ltd. (Dengyun), Zhaoqing Honda Foundry Col, Ltd. (Honda), Guangdong Hongtu Technology (Holdings) Co., Ltd. (Hongtu), Guangdong Sihui ShiLi Connecting-Rod Co., Ltd. (Shili), Guangdong Zhaoqing Power Foundry (Holding) Co., Ltd. (Power), Guangdong Hong Teo Accurate Technology Co., Ltd. (Hong Teo), Delta Aluminium Industry
Enterprises and services
Assume that there are registered enterprises in the current cloud manufacturing system, which are denoted by (Fig. 2). Enterprise () offers () different types of manufacturing services (such as design services, production services, and processing services), which are selected randomly from total types of services in the entire cloud manufacturing system. The th () type of service is denoted by . The following attributes of , including
A concrete example for the multi-task scheduling model
Fig. 4 presents a multi-task scheduling example. In this example, there are 10 enterprises, i.e. to . Each enterprise provides 4 different types of services, and each type of service is provided by 4 enterprises (the vertical axis). The letters (from a to j) in the parentheses for indicate the type of the services. There are 10 tasks (numbering from 1 to 10) in the scheduling scenario shown in Fig. 4, and each task has four subtasks with a sequence structure. In Fig. 4. tasks are
Simulation setup
The default simulation parameters are shown in Table 8. The interval parameters such as follow the uniform distribution. In Table 8, only the atomic variables are shown and the composite variables such as and can be derived from these atomic variables according to their definitions. For example, , , and , then .
Simulation results
In the following, simulation results without and with time constraint are presented for the different
Conclusion and discussions
In this paper, we proposed a workload-based multi-task scheduling model for cloud manufacturing. Based on this model, we explored the effects of two different workload-based task scheduling orders on system performance. In this model, we proposed new task workload and service modelling methods, which incorporate new ingredients such as service quantity, service efficiency, task workload, and enterprise capacity. Logistics as an important factor that can greatly affect task scheduling has also
Acknowledgements
This work was supported by the National Natural Science Foundation of China (NSFC) under Grant Nos. 61374199 and 51475032, Natural Science Foundation of Beijing, China under Grant No. 4142031, China Postdoctoral Science Foundation under Grant Nos. 2012M520139 and 2013T60052, the Fundamental Research Funds for the Central Universities under Grant No. JB140410, and the International Postdoctoral Exchange Fellowship Program under Grant No. 20140029. Special acknowledgement is given to Guangdong
References (29)
From cloud computing to cloud manufacturing
Robot. Comput.-Integr. Manuf.
(2012)- et al.
A task scheduling algorithm based on QoS-driven in Cloud Computing
Procedia Comput. Sci.
(2013) - et al.
Towards energy and resource efficient manufacturing: a processes and systems approach
CIRP Ann.-Manuf. Technol.
(2012) - et al.
Cloud manufacturing: a new service-oriented networked manufacturing model
Comput. Integr. Manuf. Syst.
(2010) - et al.
Cloud manufacturing: a new manufacturing paradigm
Enterp. Inf. Syst.
(2014) - et al.
Cloud manufacturing – a critical review of recent development and future trends
Int. J. Comput. Integr. Manuf.
(2015) - et al.
A TQCS-based service selection and scheduling strategy in cloud manufacturing
Int. J. Adv. Manuf. Technol.
(2016) - et al.
FC-PACO-RM: a parallel method for service composition optimal-selection in cloud manufacturing system
IEEE Trans. Ind. Inform.
(2013) - et al.
Cloud manufacturing service composition based on QoS with geo-perspective transportation using an improved Artificial Bee Colony optimisation algorithm
Int. J. Prod. Res.
(2015) - et al.
Correlation-aware QoS modelling and manufacturing cloud service composition
J. Intell. Manuf.
(2015)
Batch task scheduling-oriented optimization modelling and simulation in cloud manufacturing
Int. J. Simul. Model.
Multitask oriented virtual resource integration and optimal scheduling in cloud manufacturing
J. Appl. Math.
Cited by (219)
Dynamic decision-making for knowledge-enabled distributed resource configuration in cloud manufacturing considering stochastic order arrival
2024, Robotics and Computer-Integrated ManufacturingDual-service integrated scheduling of manufacturing and logistics for multiple tasks in cloud manufacturing
2024, Expert Systems with ApplicationsCloud manufacturing architectures: State-of-art, research challenges and platforms description
2023, Journal of Industrial Information IntegrationLogistics-involved task scheduling in cloud manufacturing with offline deep reinforcement learning
2023, Journal of Industrial Information IntegrationMulti-task scheduling in cloud remanufacturing system integrating reuse, reprocessing, and replacement under quality uncertainty
2023, Journal of Manufacturing Systems