Expert systems: An application in flexible manufacturing

Flexible manufacturing systems (FMS) are needed to provide manufacturing operations with the capability to adjust, in real time, to changes in the manufacturing environment. Realization of the goals of flexible manufacturing is governed by the ability of the FMS to maintain adequate information on the factory to assist in generating scenarios from product planning to operations and performance. This leads to a view where the factory is represented as an integrated information system. To facilitate the analysis of information requirements and the design of information systems for flexible manufacturing, an expert support system (ESS) which can be used to model and study the various structures is described. This ESS uses the information cell model to build these information structures. Petri net representations of these structures and their interactions are then constructed. The ESS may now be used to exercise these models and study their performance using time and cost measures.     

Real-time feedback scheduling of flexible manufacturing systems

A genetic manufacturing environment is considered. The emphasis is on small-lot, discrete, and asynchronous type of manufacturing systems rather than high volume and continuous type. Two classes of scheduling policies are proposed to render the machine stable. The policies are of feedback type. The decision is made in real-time and on-line.     

不完全柔性制造系统的最优控制

考虑具有随机需求的不完全柔性制造系统的最优控制，系统在各种产品间的切换时间是不可忽略的。运用马尔可夫最优决策过程归纳方法，导出机器服务率的最优控制策略。通过分析最优值函数的性质，证明最优策略具有简单的阈值结构，从而可得到次优生产策略--阈值控制策略。     

基于Petri网的柔性制造系统无死锁遗传调度算法

对以最小化加工时间为目标的柔性制造系统无死锁调度问题, 提出了一种遗传调度算法. 算法考虑到同类工件具有预先确定的相同加工路径, 而各工序的处理时间与工件有关. 用Petri网对工序和资源分配进行逻辑建模,利用遗传算法, 采用工序自然编码方式, 基于系统的最佳避免死锁Petri网控制器, 检测染色体的可行性, 修复不可行染色体使其对应的调度满足资源约束和无死锁控制约束, 从而保证算法所利用的所有染色体都对应系统的可行调度. 仿真结果表明了算法的可行性和有效性.     

There is no optimal routing policy for the torus

A routing policy is the method used to select a specific output channel for a message from among a number of acceptable output channels. An optimal routing policy is a policy that maximizes the probability of a message reaching its destination without delays. Optimal routing policies have been proposed for several regular networks, including the mesh and the hypercube. An open problem in interconnection network research has been the identification of an optimal routing policy for the torus. In this paper, we show that there is no optimal routing policy for the torus. Our result is demonstrated by presenting a detailed example in which the best choice of output channel is dependent on the probability of each channel being available. This result settles, in the negative, a conjecture by Wu concerning an optimal routing policy for the torus.     

Autonomy-subnet based structural synthesis and liveness guarantying policy of Petri net model of flexible manufacturing system

In this paper, an autonomous subnet based structural methodology forbottom-up synthesis of Petri Nets for Flexible Manufacturing Systems is proposed.Furthermore, the theoretical analysis of the model constructed by the method is carried byintensively using model's structural information, such as invariants, siphons, etc.. As aresult, the analysis leads us to draw the general conclusion that the model obtained isconservativeness and thus bound, and characterize its liveness in terms of zero-markingminimal siphons. It is based on model's structural information that distinguishes ourmethod from others. in line of this thought, a liveness guarantying policy for the obtainedmodel is proposed. Some control subnets are merged into the original model according tothe proposed synthesis rules in this paper to ensure that no minimal siphons are emptiedin any state, therefore the liveness is guaranteed. As a result, a live, conservative andrevertible Petri Nets is obtained. A practical example is also presented to     

An information system for computer-integrated manufacturing systems

One of the main problems when implementing the computer-integrated manufacturing (CIM) concept concerns information integration. In order to support information integration, an information system provided with suitable data models is required. In this paper an information system is presented, which fulfils the requirements for an appropriate information management in CIM. For this purpose, an EXPRESS-oriented information system has been built on top of a commercial relational DBMS. The conceptual model of the information is built in EXPRESS and then parsed and translated to the corresponding relational constructs. A data access interface has been implemented which allows EXPRESS-oriented data manipulation. The information system was developed inside ESPRIT project No. 2202 CIM-PLATO “CIM System Planning Toolbox” to integrate the information used by the tools developed for flexible manufacturing system planning.     

Applying Gaia and AUML for the development of multiagent‐based control software for flexible manufacturing systems: addressing methodological and implementation issues

In this article, we present the development of a simple multiagent‐based system for the control of a flexible manufacturing system. We followed the stages of a methodology specially conceived for the development of agent‐based system, which is an integration of the classical methodology for agent‐oriented analysis and design Gaia, and AUML (Agent‐Unified Modeling Language). We adopted as study case the CIMUBB Laboratory at the University of Bio‐Bio, which has a flexible manufacturing system including three flexible manufacturing cells interconnected by a conveyor belt. In the analysis stage, we identified roles involved, and we design models representing roles and protocols. In the design stage, we applied Gaia agent, services, and acquaintance models from Gaia, and we complemented with AUML as the adopted methodology suggests. With the developed models, we constructed a fully functional system where each agent was built as an independent process tree. Agents communicate by passing messages through the Ethernet network with socket interfaces. Various tests executed in our laboratory scale manufacturing system show the effectiveness of our implementation. Copyright © 2015 John Wiley & Sons, Ltd.     

应用控制变迁的柔性制造系统死锁控制策略

不同于目前许多文献中基于添加控制库所的死锁预防策略,本文提出了控制变迁方程(CTE)的概念和相应的基于添加控制变迁(CT)的死锁控制策略(DCP).通过分析存在死锁的原网(N0, M0)的可达图(RG),该DCP求解出所有死锁标识(DM).基于CTE,构造出所需的控制变迁.然后,对每个DM添加相应的CT,进而消除了原网(N_0, M_0)中的死锁标识,得到了活性受控网系统(N~?, M~?).通过理论分析和相关算例的应用,该DCP的正确性和有效性得到了验证.此外,该DCP获取的活性受控网系统(N~?, M~?)可达数目与原网(N_0, M_0)是相同的,即最大可达数(MRN).     

Expert systems for manufacturing cell simulation and design

Computer Integrated Manufacturing (CIM) is approached by means of the application of Computer Aided Design (CAD), Computer Aided Manufacturing (CAM) and other CA techniques, methods and programs/program systems. These programs are often implemented as knowledge-based, or expert systems and in this way they became typical examples of engineering application of artificial intelligence. The production task of CIM systems is solved by using flexible manufacturing systems (FMSs). FMSs built up from smaller, complex units, i.e. from flexible manufacturing cells (FMCs) have several advantages. The design and the operation of manufacturing systems need new, sophisticated methods to utilize all the embedded benefits of the sophisticated and expensive elements installed for production purposes. New methods like knowledge processing technology, cooperative problem-solving techniques, etc., offer wide possibilities to design more reasonable systems. This paper describes prototype expert systems that make use of different knowledge-based tools and techniques to design (configure, reconfigure) and simulate manufacturing cells, taking into consideration technological plans and other relevant information.     

Modelling manufacturing systems: A birth-death approach

An approximation method for modelling a manufacturing system is introduced. The system is considered as a queueing network, where each queue is limited in size, and interarrival and processing times are exponentially distributed. The birth-death approach is considered and an approximation method to reduce the dimension of the model is developed. The results are the marginal probability distribution of the number of units in each queue; other performance indices, such as mean queue lengths, utilizations of the working stations, and throughput can be easily obtained. The general procedure is applied to model, for example, queues in tandem, a split node, and a more complex network of queues. Simulation and, when possible, comparison with the exact solution show an acceptable error level of the proposed method.     

Developing control and integration software for flexible manufacturing systems

The slow growth of computer-integrated manufacturing is attributed to the complexity of designing and implementing their control and integration software. This article expands on a methodology for designing and implementing this software that was introduced in [16]. The goal of this methodology is to build flexible and resuable control and integration software for computer-integrated manufacturing systems. It hinges upon the concepts of software/hardware components, their assemblages, a distributed common language environment, formal models, and generic controllers. Major sources of flexibility are obtained by decoupling process plan models from the model of the factory floor and by using a generic controller. Reusability is achieved by building selfcontained software/hardware components with general, possibly parametrized, interfaces. The interplay between simulated and actual hardware internals of software/hardware components is used as the basis of a testing strategy that performs off-line simulation followed by on-line testing.The methodology has been applied in designing and implementing the control and integration software of an actual Prismatic Machining Cell. The article also reports on the details of this implementation.The names of the authors appear in alphabetical order.     

面向柔性制造车间的多目标RFID网络规划方法

针对具有多种覆盖需求的柔性制造车间RFID网络规划问题,以部署成本、阅读器干扰与阅读器效能为多优化目标,提出一种分层聚类、冗余消减与梯度下降方法集成的RFID网络规划方法,采用分层聚类算法确定RFID初始数量与部署位置,采用冗余阅读器消减算法优化RFID数量,采用梯度下降算法优化RFID部署位置从而实现网络规划多目标优化。实验研究表明,提出的方法在多目标综合性能方面优于传统基于分层聚类方法、遗传算法、粒子群与冗余消减混合方法的RFID网络规划方法,验证了该方法的有效性。     

On structural minimality of optimal supervisors for flexible manufacturing systems

This paper develops a place invariant based deadlock prevention method to obtain an optimal, i.e., maximally permissive, liveness-enforcing Petri net supervisor with a minimal supervisory structure that means the minimal number of control places. Maximal permissiveness can be achieved by designing place invariants that make all legal markings reachable while all first-met bad markings unreachable. An integer linear programming problem is formulated to compute all place invariants and its objective function minimizes the number of place invariants, aiming to yield a minimal supervisory structure. Importantly, we develop a technique to greatly improve the efficiency of the proposed method by reducing the number of constraints and variables in the integer linear programming problem under consideration. A number of examples from the literature are used to illustrate the proposed approaches.     

基于变迁覆盖的制造系统死锁控制策略

基于系统Petri网模型,研究柔性制造系统的死锁控制问题.论文利用变迁覆盖为系统设计活性控制器.变迁覆盖是由一组极大完备资源变迁回路组成的集合,其变迁集覆盖了Petri网中所有极大完备资源变迁回路的变迁集.验证变迁覆盖的有效性,然后仅对有效变迁覆盖中的极大完备资源变迁回路添加控制位置,就得到系统的活性受控Petri网.这种受控Petri网包含的控制位置个数少,从而结构相对简单.最后通过一个例子说明了所提出的死锁控制策略的构成与特点.     

Hybrid heuristic search approach for deadlock-free scheduling of flexible manufacturing systems using Petri nets

Based on the Petri net models of flexible manufacturing systems (FMSs), this paper focuses on deadlock-free scheduling problem with the objective of minimizing the makespan. Two hybrid heuristic search algorithms for solving such scheduling problems of FMSs are proposed. To avoid deadlocks, the deadlock control policy is embedded into heuristic search strategies. The proposed algorithms combine the heuristic best-first strategy with the controlled backtracking strategy based on the execution of the Petri nets. The scheduling problem is transformed into a heuristic search problem in the reachability graph of the Petri net, and a schedule is a transition sequence from the initial marking to the final marking in the reachability graph. By using the one-step look-ahead method in the deadlock control policy, the safety of a state in the reachability graph is checked, and hence, deadlock is avoided. Experimental results are provided and indicate the effectiveness of the proposed hybrid heuristic search algorithms in solving deadlock-free scheduling problems of FMSs. Especially, the comparison against previous work shows that both new algorithms are promising in terms of solution quality and computing times.     

A strategic investment model for phased implementation of flexible manufacturing systems

This paper describes a set of strategic, tactical, and operational models that can be used to analyze the phased implementation of flexible technology in a manufacturing system. The strategic model represents capital investment decisions, the tactical model represents aggregate production decisions, and the operational model represents the functional form of the production costs. Learning effects and other nonlinearities (such as setup costs, economies of scale, and congestion effects) can be quantified in the models. An efficient solution procedure (using dynamic programming and minimum cost network flow optimization) is described. The efficiency of the solution method permits a host of 'what-if' scenarios to be examined by the decision maker.     

Genetic algorithms for match-up rescheduling of the flexible manufacturing systems

Scheduling plays a vital role in ensuring the effectiveness of the production control of a flexible manufacturing system (FMS). The scheduling problem in FMS is considered to be dynamic in its nature as new orders may arrive every day. The new orders need to be integrated with the existing production schedule immediately without disturbing the performance and the stability of existing schedule. Most FMS scheduling methods reported in the literature address the static FMS scheduling problems. In this paper, rescheduling methods based on genetic algorithms are described to address arrivals of new orders. This study proposes genetic algorithms for match-up rescheduling with non-reshuffle and reshuffle strategies which accommodate new orders by manipulating the available idle times on machines and by resequencing operations, respectively. The basic idea of the match-up approach is to modify only a part of the initial schedule and to develop genetic algorithms (GAs) to generate a solution within the rescheduling horizon in such a way that both the stability and performance of the shop floor are kept. The proposed non-reshuffle and reshuffle strategies have been evaluated and the results have been compared with the total-rescheduling method.