Multi-agent based real-time production scheduling method for radio frequency identification enabled ubiquitous shopfloor environment

The lack of timely feedback shopfloor information during manufacturing execution stage leads to significant difficulties in achieving real-time production scheduling. To address this problem, an overall architecture of multi-agent based real-time production scheduling is presented to close the loop of production planning and control. Several contributions are significant. Firstly, wireless devices such as radio frequency identification (RFID) are deployed into value-adding points in a ubiquitous shopfloor environment to form Machine Agent for the collection and processing of real-time shopfloor data. Secondly, Capability Evaluation Agent is designed to optimally assign the tasks to the involved machines at the process planning stage based on the real-time utilization ration of each machine. The third contribution is a Real-time Scheduling Agent for manufacturing tasks scheduling/re-scheduling strategy and methods according to the real-time feedback. Fourthly, a Process Monitor Agent model is designed for tracking and tracing the manufacturing execution based on a critical event structure. Finally, a case is used to demonstrate the proposed multi-agent based real-time production scheduling models and methods.     

Application of reinforcement learning for agent-based production scheduling

Reinforcement learning (RL) has received some attention in recent years from agent-based researchers because it deals with the problem of how an autonomous agent can learn to select proper actions for achieving its goals through interacting with its environment. Although there have been several successful examples demonstrating the usefulness of RL, its application to manufacturing systems has not been fully explored yet. In this paper, Q-learning, a popular RL algorithm, is applied to a single machine dispatching rule selection problem. This paper investigates the application potential of Q-learning, a widely used RL algorithm to a dispatching rule selection problem on a single machine to determine if it can be used to enable a single machine agent to learn commonly accepted dispatching rules for three example cases in which the best dispatching rules have been previously defined. This study provided encouraging results that show the potential of RL for application to agent-based production scheduling.     

Routing distributions and their impact on dispatch rules

Motivated by the job-shop production process of our industry partner, we examine dispatching rules effects on two key performance indicators (KPIs) – job lateness and the percentage of late jobs. In the literature, authors use the uniform distribution to generate random job shop data. In addition to our discussion on dispatching rules, we propose an alternative idea for random job shop data, the routing distribution, and we compare dispatching rules performance using KPI frontiers under different routing distributions. We show that using their current dispatch rule, earliest operation due date (EODD), the industry partner is never worse off, even as their job-shop’s operational environment changes. We further show that using multiple dispatch rules across several job-shop departments does improve a job-shop’s performance on the KPIs, though the improvement is small and in some cases may not be statistically significant. In addition, we find that EODD is one of several dispatching rule which consistently lie on the KPI frontier for different job routing distributions. We find that dispatching rule performance is greatly affected by the routing distribution of the job-shop where the rules are employed. Lastly, we leave the readers with some insight into determining which dispatch rules and routing distributions should be considered for different job shops.     

Distributed Artificial Intelligence in manufacturing systems control

This research uses a Distributed Artificial Intelligence (DAI) framework to efficiently utilize the infrastructure available for process planning in a batch processing PWB assembly facility. The DAI approach decomposes the entire production control task into several sub-tasks. Then, the sub-tasks are implemented by the basic elements of the DAI system called ‘intelligent agents’. By working collectively, the intelligent agents of the DAI system can arrive at a solution for the problem. The DAI system initially proposes all possible solutions generated by the intelligent agents. Then, a fuzzy coordination technique is utilized to evaluate the solutions and to find the most appropriate one for shopfloor implementation. Using inputs such as the short-term production plan, design data, shopfloor observation data, and CAD information, the DAI system provides applicable production plans with ranks for the feasibility of current assembly activities.     

Rapid Modeling and Discovery of Priority Dispatching Rules: An Autonomous Learning Approach

Priority-dispatching rules have been studied for many decades, and they form the backbone of much industrial scheduling practice. Developing new dispatching rules for a given environment, however, is usually a tedious process involving implementing different rules in a simulation model of the facility under study and evaluating the rule through extensive simulation experiments. In this research, an innovative approach is presented, which is capable of automatically discovering effective dispatching rules. This is a significant step beyond current applications of artificial intelligence to production scheduling, which are mainly based on learning to select a given rule from among a number of candidates rather than identifying new and potentially more effective rules. The proposed approach is evaluated in a variety of single machine environments, and discovers rules that are competitive with those in the literature, which are the results of decades of research.     

A multi-agent based approach to dynamic scheduling of machines and automated guided vehicles in manufacturing systems

In real manufacturing environments, the control of system elements such as automated guided vehicles has some difficulties when planning operations dynamically. Multi agent-based systems, a newly maturing area of distributed artificial intelligence, provide some effective mechanisms for the management of such dynamic operations in manufacturing environments. This paper proposes a multi-agent based scheduling approach for automated guided vehicles and machines within a manufacturing system. The proposed multi-agent based approach works under a real-time environment and generates feasible schedules using negotiation/bidding mechanisms between agents. This approach is tested on off-line scheduling problems from the literature. The results show that our approach is capable of generating good schedules in real time comparable with the optimization algorithms and the frequently used dispatching rules.     

A two-level advanced production planning and scheduling model for RFID-enabled ubiquitous manufacturing

Radio frequency identification (RFID) technology has been used in manufacturing industries to create a RFID-enabled ubiquitous environment, in where ultimate real-time advanced production planning and scheduling (APPS) will be achieved with the goal of collective intelligence. A particular focus has been placed upon using the vast amount of RFID production shop floor data to obtain more precise and reasonable estimates of APPS parameters such as the arrival of customer orders and standard operation times (SOTs). The resulting APPS model is based on hierarchical production decision-making principle to formulate planning and scheduling levels. A RFID-event driven mechanism is adopted to integrate these two levels for collective intelligence. A heuristic approach using a set of rules is utilized to solve the problem. The model is tested through four dimensions, including the impact of rule sequences on decisions, evaluation of released strategy to control the amount of production order from planning to scheduling, comparison with another model and practical operations, as well as model robustness. Two key findings are observed. First, release strategy based on the RFID-enabled real-time information is efficient and effective to reduce the total tardiness by 44.46% averagely. Second, it is observed that the model has the immune ability on disturbances like defects. However, as the increasing of the problem size, the model robustness against emergency orders becomes weak; while, the resistance to machine breakdown is strong oppositely. Findings and observations are summarized into a number of managerial implications for guiding associated end-users for purchasing collective intelligence in practice.     

Dynamic routing strategies for JIT production in hybrid flow shops

In production processes, just-in-time (JIT) completion of jobs helps reduce both the inventory and late delivery of finished products. Previous research which aims to achieve JIT job completion mainly worked on static scheduling problems, in which all jobs are available from time zero or the available time of each job is known beforehand. In contrast, dynamic scheduling problems which involve continual arrival of new jobs are not much researched and dispatching rules remain the most frequently used method for such problems. However, dispatching rules are not high-performing for the JIT objective. This study proposes several routing strategies which can help dispatching rules realize JIT completion for jobs arriving dynamically in hybrid flow shops. The routing strategies are based on distributed computing which makes realtime forecast of completion times of unfinished jobs. The advantages include short computing time, quick response and robustness against disturbance. Computer simulations show that the performance of dispatching rules combined with the proposed routing strategies is significantly higher than that of dispatching rules only and that of dispatching rules combined with the previous routing methods.     

Application of intelligent data management in resource allocation for effective operation of manufacturing systems

Resource allocation has been a critical issue in manufacturing. This paper presents an intelligent data management induced resource allocation system (RAS) which aims at providing effective and timely decision making for resource allocation. This sophisticated system is comprised of product materials, people, information, control and supporting function for the effectiveness in production. The said system incorporates a Database Management System (DBMS) and fuzzy logic to analyze data for intelligent decision making, and Radio Frequency Identification (RFID) for result verification. Numerical data from diverse sources are managed in the DBMS and used for resource allocation determination by using fuzzy logic. The output, representing the essential resources level for production, is then verified with reference to the resource utilization status captured by RFID. The effectiveness of the developed system is verified with a case study carried out in a Hong Kong-based garment manufacturing company. Results show that data gathering before resource allocation determination is more efficient with the use of developed system where the resource allocation decision parameters in the centralize database are effectively determined by using fuzzy logic. Decision makers such as production managers are allowed to determine resource allocation in a standardized approach in a more efficient way. The system also incorporates RFID with Artificial Intelligence techniques for result verification and knowledge refinement. Therefore, fuzzy logic results of resource allocation can be more responsive and adaptive to the actual production situation by refining the fuzzy rules with reference to the RFID-captured data.     

动态车间作业调度问题中调度规则算法研究综述

调度规则是解决实际生产中的动态车间作业调度问题的有效方法,但它一般只在特定调度环境下性能较好,当环境发生变化时,就需要进行实时选择和评价。对调度规则的实时选择和评价方法进行综述,以研究实际生产中动态车间的实时调度问题。对调度规则的发展、分类以及特点进行了概述,并对调度规则的选择和评价方法进行了总结。详细介绍了调度规则的选择方法,包括使用较多的稳态仿真方法和表现较好的人工智能方法,并给出了仿真方法、专家系统、机器学习方法以及人工神经网络方法,用于调度规则的选择时所取得的研究成果和结论。此外,还介绍了调度规则的评价指标及评价方法。最后针对调度规则存在的不足,指出了未来的研究方向。     

A neural network job-shop scheduler

This paper focuses on the development of a neural network (NN) scheduler for scheduling job-shops. In this hybrid intelligent system, genetic algorithms (GA) are used to generate optimal schedules to a known benchmark problem. In each optimal solution, every individually scheduled operation of a job is treated as a decision which contains knowledge. Each decision is modeled as a function of a set of job characteristics (e.g., processing time), which are divided into classes using domain knowledge from common dispatching rules (e.g., shortest processing time). A NN is used to capture the predictive knowledge regarding the assignment of operation’s position in a sequence. The trained NN could successfully replicate the performance of the GA on the benchmark problem. The developed NN scheduler was then tested against the GA, Attribute-Oriented Induction data mining methodology and common dispatching rules on a test set of randomly generated problems. The better performance of the NN scheduler on the test problem set compared to other methods proves the feasibility of NN-based scheduling. The scalability of the NN scheduler on larger problem sizes was also found to be satisfactory in replicating the performance of the GA.     

Hybrid genetic algorithmsfor a multiple-objective scheduling problem

This paper describes the characteristics of two hybrid genetic algorithms (GAs) for generating allocation and sequencing of production lots in a flow-shop environment based on a non-linear, multi-criteria objective function. Both GAs are used as search techniques: in the first model the task of the GA is to allocate and sequence the jobs; in the second model, the GA is combined with a dispatching rule (Earliest Due Date, EDD) thus limiting its task only on the allocation of the jobs. Both GAs are characterized by a dynamic population size with dynamic birth rate, as well as by multiple-operator reproduction criteria and by adaptive crossover and mutation rates. A discrete-event simulation model has been used in order to evaluate the performances of the tentative schedules. The proposed algorithms have been subsequently compared with a classical branch and bound method.     

A Train Dispatching Model Under a Stochastic Environment: Stable Train Routing Constraints and Reformulation

After major capacity breakdown(s) on a railway network, train dispatchers need to generate appropriate dispatching plans to recover the impacted train schedule from perturbations and minimize the expected total train delay time under stochastic scenarios. In this paper, we propose a cumulative flow variables-based integer programming model for dispatching trains under a stochastic environment on a general railway network. Stable Train Routing (STR) constraints are introduced to ensure that trains traverse on the same route across different capacity breakdown scenarios, which are further reformulated to equivalent linear inequality constraints. Track occupancy and safety headways are modelled as side constraints which are dualized through a proposed Lagrangian relaxation solution framework. The original complex train dispatching problem is then decomposed to a set of single-train and single-scenario optimization subproblems. For each subproblem, a standard label correcting algorithm is embedded for finding the time dependent least cost path on a space-time network. The resulting dual solutions can be transformed to feasible solutions through priority rules. Numerical experiments are conducted to demonstrate the efficiency and effectiveness of the proposed solution approach.     

智能制造系统基于数据驱动的车间实时调度

智能制造系统采用大量先进的信息技术,为车间实时调度提供技术基础.各类信息技术在生产制造过程中的广泛应用使得制造系统积累了大量与生产调度相关的数据,因此,通过利用历史生产调度数据和智能装备收集到的实时生产数据,建立基于数据驱动的生产实时调度方法成为新型制造环境下实现高效调度的新思路.针对智能制造环境下的混合流水车间实时调度问题,提出基于BP神经网络的数据驱动的实时调度方法,从历史近优的调度方案中提取用于调度知识挖掘的样本数据,通过BP神经网络训练学习获取生产系统状态与调度规则的映射关系网络,并将其应用于生产在线实时调度.数值实验表明,所提出的方法优于固定单一调度规则,在不同的调度性能指标下其效果均稳定且良好.     

RFID生产线监控及调度管理系统

针对传统离散制造企业数据采集实时性差、生产管理信息化程度低等问题,引入一种基于RFID技术的监控管理系统。该系统采用RFID技术与软硬件中间件、CAN总线通信及实时数据平台等技术相结合的方法,实现了离散生产过程的实时数据采集、可视化监控和优化调度管理。该系统已成功应用于几家服装企业,显著提高了企业的生产效率、产品质量和管理信息化程度,这对其他离散制造领域具有推广应用价值。     

RFID-based Production Data Analysis in an IoT-enabled Smart Job-shop

Under industry 4.0, internet of things (IoT), especially radio frequency identification (RFID) technology, has been widely applied in manufacturing environment. This technology can bring convenience to production control and production transparency. Meanwhile, it generates increasing production data that are sometimes discrete, uncorrelated, and hard-to-use. Thus, an efficient analysis method is needed to utilize the invaluable data. This work provides an RFID-based production data analysis method for production control in IoT-enabled smart job-shops. The physical configuration and operation logic of IoT-enabled smart job-shop production are firstly described. Based on that, an RFID-based production data model is built to formalize and correlate the heterogeneous production data. Then, an eventdriven RFID-based production data analysis method is proposed to construct the RFID events and judge the process command execution. Furthermore, a near big data approach is used to excavate hidden information and knowledge from the historical production data. A demonstrative case is studied to verify the feasibility of the proposed model and methods. It is expected that our work will provide a different insight into the RFIDbased production data analysis.      

Holonic supply chain: A study from family-based manufacturing perspective

In the contemporary business environment, to adhere to the need of the customers, caused the shift from mass production to mass-customization. This necessitates the supply chain (SC) to be effective flexible. The purpose of this paper is to seek flexibility through adoption of family-based dispatching rules under the influence of inventory system implemented at downstream echelons of an industrial supply chain network. We compared the family-based dispatching rules in existing literature under the purview of inventory system and information sharing within a supply chain network. The dispatching rules are compared for Average Flow Time performance, which is averaged over the three product families. The performance is measured using extensive discrete event simulation process. Given the various inventory related operational factors at downstream echelons, the present paper highlights the importance of strategically adopting appropriate family-based dispatching rule at the manufacturing end. In the environment of mass customization, it becomes imperative to adopt the family-based dispatching rule from the system wide SC perspective. This warrants the application of intra as well as inter-echelon information coordination. The holonic paradigm emerges in this research stream, amidst the holistic approach and the vital systemic approach. The present research shows its novelty in triplet. Firstly, it provides leverage to manager to strategically adopting a dispatching rule from the inventory system perspective. Secondly, the findings provide direction for the attenuation of adverse impact accruing from demand amplification (bullwhip effect) in the form of inventory levels by appropriately adopting family-based dispatching rule. Thirdly, the information environment is conceptualized under the paradigm of Koestler’s holonic theory.     

Heuristic Method for Job Shop Scheduling Problem with Combination Processing

A job shop scheduling problem with a combination processing in complex production environment is proposed. Based on the defining of non-elastic combination processing relativity and virtual process,the problem can be simplified and transformed to a traditional one. On the basis of the dispatching rules select engine and considered factors of complex production environment,a heuristic method is designed. The algorithm has been applied to a mould enterprise in Shenzhen for half a year. The practice showed tha...     

APS 9: an improved adaptive population-based simplex method for real-world engineering optimization problems

The adaptive population-based simplex (APS) algorithm is a recently-proposed optimization method for solving continuous optimization problems. In this paper, a new variant of APS, referred to as APS 9, is proposed to solve engineering optimization problems. APS 9 still follows the main structure of APS where three strategies (i.e., reflection, contraction and local search) can be used to improve the population of solutions. However, the three strategies have been significantly modified and the rules for applying them have been revised. A stagnation detection mechanism and duplicates removal step have been added. The proposed method is compared with the winners of the IEEE CEC 2005 and CEC 2011 competitions on the 22 CEC 2011 problems. The results show the superiority of APS 9 compared to the other two methods. Moreover, APS 9 has been compared with two recent optimization methods on the same test bed. The limitations of the CEC 2011 competition are also discussed and new rules that are more engineering-friendly are proposed.     

Dispatching heuristics for the single machine early/tardy scheduling problem with job-independent penalties

In this paper, we consider the single machine earliness/tardiness scheduling problem with job-independent penalties, and no machine idle time. Several dispatching heuristics are proposed, and their performance is analysed on a wide range of instances. The heuristics include simple scheduling rules, as well as a procedure that takes advantage of the strengths of each of those rules. We also consider early/tardy dispatching procedures, and a heuristic method based on existing adjacent precedence conditions. An improvement procedure that can be used to improve the schedules generated by the heuristics is also proposed.

The computational tests show that the best results are given by the early/tardy dispatching rules. These heuristics are also quite fast, and are capable of quickly solving even very large instances. The use of the improvement procedure is recommended, since it improves the solution quality, with little additional computational effort.