A feedforward multi-layer neural network for machine cell formation in computer integrated manufacturing

Neural-network applications have been one of the better alternatives either for simulating massive data in parallel, or embedding human subjective decisions into existing quantitative models, thereby spawning a qualitative model. This paper introduces a linear classifier with a classical feedforward neural network in forming machine cells or groups for Computer Integrated Manufacturing. The proposed method, through experiment, has been proven to outperform conventional methods such as Part Family Analysis (PFA) and BLOCPLAN, among others. A single-layer perceptron, along with multi-layer feedforward network where applicable, have been employed in forming the part families. The underlying philosophy is the Group Technology (GT). The developed models and algorithms are illustrated with a numerical example.     

Comparison of visualization of optimal clustering using self-organizing map and growing hierarchical self-organizing map in cellular manufacturing system

The present research deals with the cell formation problem (CFP) of cellular manufacturing system which is a NP-hard problem thus, the development of optimum machine-part cell formation algorithms has always been the primary attraction in the design of cellular manufacturing system. In this proposed work, the self-organizing map (SOM) approach has been used which is able to project data from a high-dimensional space to a low-dimensional space so it is considered a visualized approach for explaining a complicated CFP data set. However, for a large data set with a high dimensionality, a traditional flat SOM seems difficult to further explain the concepts inside the clusters. We propose one such possible solution for a large CFP data set by using the SOM in a hierarchical manner known as growing hierarchical self-organizing map (GHSOM). In the present work, the two novel contributions using GHSOM are: the choice of optimum architecture through the minimum pattern units extracted at layer 1 for the respective threshold values and selection. Furthermore, the experimental results clearly indicated that the machine-part visual clustering using GHSOM can be successfully applied in identifying a cohesive set of part family that is processed by a machine group. Computational experience specifically with the proposed GHSOM algorithm, on a set of 15 CFP problems from the literature, has shown that it performs remarkably well. The GHSOM algorithm obtained solutions that are at least as good as the ones found the literature. For 75% of the cell formation problems, the GHSOM algorithm improved the goodness of cell formation through GTE performance measure using SOM as well as best one from the literature, in some cases by as much as more than 12.81% (GTE). Thus, comparing the results of the experiment in this paper with the SOM and GHSOM using the paired t-test it has been revealed that the GHSOM approach performed better than the SOM approach so far the group technology efficiency (GTE) measures of performance of the goodness of cell formation is concerned.     

Neural network-based design of cellular manufacturing systems

A neural network based on a competitive learning rule, when trained with the part machine incidence matrix of a large number of parts, classifies the parts and machines into part families and machine cells, respectively. This classification compares well with the classical clustering techniques. The steady state values of the activations and interconnecting strengths enable easier identification of the part families, machine cells, overlapping parts and bottleneck machines. Neural networks are mostly applied by treating them as a blackbox, i.e. the interaction with the environment and the information acquisition and retrieval occurs at the input and the output level of the network. This paper presents an approach where knowledge is extracted from the external and internal structure of the neural network.     

Design of virtual manufacturing cells: a mathematical programming approach

In this paper, a new type of virtual cellular manufacturing (CM) system is considered, and a multi-objective design procedure is developed for designing such cells in real time. Retaining the functional layout, virtual cells are addressed as temporary groupings of machines, jobs and workers to realize the benefits of CM. The virtual cells are created periodically, for instance every week or every month, depending on changes in demand volumes and mix, as new jobs accumulate during a planning period. The procedure includes labor grouping considerations in addition to part-machine grouping. The procedure is based on interactive goal programming methods. Factors such as capacity constraints, cell size restrictions, minimization of load imbalances, minimization of inter-cell movements of parts, provision of flexibility, etc. are considered. In labor grouping, the functionally specialized labor pools are partitioned and regrouped into virtual cells. Factors such as ensuring balanced loads for workers, minimization of inter-cell movements of workers, providing adequate levels of labor flexibility, etc. are considered in a pragmatic manner.     

Structured intelligence for self-organizing manufacturing systems

This paper deals with fuzzy scheduling and path planning problems by genetic algorithms. We have proposed a self-organizing manufacturing system (SOMS) that is composed of a number of autonomous modules. Each module decides output through interaction with other modules, but the module does not share complete information concerning other modules in the SOMS. Therefore, we require structured intelligence as a whole system. In this paper, we consider a manufacturing line composed of machining centres and conveyor units. The manufacturing procedure can be divided into a sequence of three modules: (a) tool locating module, (b) scheduling module, and (c) path planning module. The tool locating problems have been already solved. In this paper, we first solve the scheduling problem as global preplanning. Here we assume that the processing time is not constant, because some delay may occur in the machining centres. We therefore apply fuzzy theory to represent incomplete information abou t the machining time. We solve the fuzzy scheduling problem with a genetic algorithm. After global preplanning, the path planning module transports materials and products. Next, the scheduling module acquires the actual processing time of each machining centre. Based on the processing time, the schedule module generates a fuzzy number for the processing time. We discuss the effectiveness of the proposed method through the computer simulation results.     

Capability-based virtual cellular manufacturing systems formation in dual-resource constrained settings using Tabu Search

Formation of Virtual Cellular Manufacturing Systems (VCMSs), as one of the main applications of Group Technology (GT), by presentation of unique and shared capability boundaries of machine tools through defining Resource Elements (REs) creates a good solution for Capability-Based VCMSs (CBVCMSs), which increases opportunities to create systems with more efficient utilizations. Considering workers as the second important resources in Dual-Resource Constraint (DRC) settings makes this problem more serious and critical to research because, in reality, jobs cannot be processed if workers, machines, or both are not available. This paper attempts to form CBVCMSs with DRC settings using a multi-objective mathematical model with a Goal Programming (GP) approach. Using the developed model, parts, machines, and workers are grouped and assigned to the generated virtual cells at the same time. The proposed model is solved through a multi-objective Tabu Search (TS) algorithm to find optimum or near-to-optimum solutions. The validity of the developed model is illustrated by two numerical examples taken from the literature and evaluated through comparing the performance of the CBVCMSs and the original classical CMSs in the System Capacity Utilization (SCU) point of view.     

A CORBA-based agent-driven design for distributed intelligent manufacturing systems

An agent-oriented methodology is presented for representation, acquisition, and processing of manufacturing knowledge along with analysis and modeling of an intelligent manufacturing system (IMS). An intelligent manufacturing system adopts heterarchical and collaborative control as its information system architecture. The behavior of the entire manufacturing system is collaboratively determined by many interacting subsystems that may have their own independent interests, values, and modes of operation. The subsystems are represented as agents. An agent's architecture and task decomposition method are presented. The agent-oriented methodology is used to analyze and model an intelligent machine cell. An intelligent machine center is considered as an autonomous, modular, reconfigurable and fault-tolerant machine tool with self-perception, decision making, and self-process planning, able to cooperate with other machines through communication. The common object request broker architecture (CORBA) distributed software control system was developed as a simple prototype. A case study illustrates an intelligent machine center.     

A methodology to incorporate product mix variations in cellular manufacturing

The identification of part families and machine groups that form the cells is a major step in the development of a cellular manufacturing system and, consequently, a large number of concepts, theories and algorithms have been proposed. One common assumption for most of these cell formation algorithms is that the product mix remains stable over a period of time. In today’s world, the market demand is being shaped by consumers resulting in a highly volatile market. This has given rise to a new class of products characterized by low volume and high variety. To incorporate product mix changes into an existing cellular manufacturing system many important issues have to be tackled. In this paper, a methodology to incorporate new parts and machines into an existing cellular manufacturing system has been presented. The objective is to fit the new parts and machines into an existing cellular manufacturing system thereby increasing machine utilization and reducing investment in new equipment.     

小波神经网络在飞控系统辨识中的应用研究

选择以sigmoid函数为基础的小波基波函数构造了一个小波神经网络,利用小波网络对复杂的飞控系统对象进行在线辨识研究,仿真结果表明小波神经网络基本满足某型飞机飞控系统在线辨识的要求。     

半导体生产线的RBF神经网络预测建模

预测模型是预测控制的基础。半导体生产过程的复杂性和随机性,使之难以建立确定性模型。提出一种新方法,利用径向基（RBF）神经网络对该过程建立预测模型。使用simul8软件对之在各种投料策略和调度策略下进行仿真并定时采样,将采样得到的数据对模型进行训练测试,结果表明该模型的预测输出与实测样本基本吻合,网络模型具有很好的泛化能力。训练后的网络可以对半导体生产线进行快速准确的预测,为预测控制和实时调度打下基础。     

A hybrid multi-objective approach based on the genetic algorithm and neural network to design an incremental cellular manufacturing system

One important issue related to the implementation of cellular manufacturing systems (CMSs) is to decide whether to convert an existing job shop into a CMS comprehensively in a single run, or in stages incrementally by forming cells one after the other, taking the advantage of the experiences of implementation. This paper presents a new multi-objective nonlinear programming model in a dynamic environment. Furthermore, a novel hybrid multi-objective approach based on the genetic algorithm and artificial neural network is proposed to solve the presented model. From the computational analyses, the proposed algorithm is found much more efficient than the fast non-dominated sorting genetic algorithm (NSGA-II) in generating Pareto optimal fronts.     

Medical image segmentation based on cellular neural network

The application of cellular neural network (CNN) has made great progress in image processing. When the selected objects extraction (SOE) CNN is applied to gray scale images, its effects depend on the choice of initial points. In this paper, we take medical images as an example to analyze this limitation. Then an improved algorithm is proposed in which we can segment any gray level objects regardless of the limitation stated above. We also use the gradient information and contour detection CNN to determine the contour and ensure the veracity of segmentation effectively. Finally, we apply the improved algorithm to tumor segmentation of the human brain MR image. The experimental results show that the algorithm is practical and effective.     

人工神经网络泛化问题研究综述*

从理论、方法（思想）和技术三个层次回顾了以往工作,讨论了模型复杂度、样本复杂度及两者之间关系的相关研究;在实际中,通过控制模型复杂度、调整样本等具体技术可以在一定程度上提高神经网络的泛化能力,但这些技术仍然存在一些问题没有解决。最后提出了对今后研究的展望。     

The modified fuzzy art and a two-stage clustering approach to cell design

This study presents a new pattern recognition neural network for clustering problems, and illustrates its use for machine cell design in group technology. The proposed algorithm involves modifications of the learning procedure and resonance test of the Fuzzy ART neural network. These modifications enable the neural network to process integer values rather than binary valued inputs or the values in the interval [0, 1], and improve the clustering performance of the neural network. A two-stage clustering approach is also developed in order to obtain an informative and intelligent decision for the problem of designing a machine cell. At the first stage, we identify the part families with very similar parts (i.e., high similarity exists in their processing requirements), and the resultant part families are input to the second stage, which forms the groups of machines. Experimental studies show that the proposed approach leads to better results in comparison with those produced by the Fuzzy ART and other similar neural network classifiers.     

基于自组织神经网络的QRS波聚类方法

为提高查看大量数据动态心电(ECG)图时的效率,将波形聚类,采用埃尔米特函数和自组织神经网络,实现了室性早搏占比高情况下的心电波形聚类算法.使用MIT-BIH心率失常数据库,利用埃尔米特函数分解QRS波形为QRS向量,将所有QRS向量输入自组织神经网络进行分类.使用特征向量元素分析聚类结果,用阳性率指标对结果进行统计,平均真阳性率为91.2％,假阳性率为1.03％,验证了基于自组织神经网络的心电聚类算法的有效性.达到了将正常心搏和室性早搏心搏聚类的目标.     

Advanced self-organizing polynomial neural network

In this paper, we introduce a concept of advanced self-organizing polynomial neural network (Adv_SOPNN). The SOPNN is a flexible neural architecture whose structure is developed through a modeling process. But the SOPNN has a fatal drawback; it cannot be constructed for nonlinear systems with few input variables. To relax this limitation of the conventional SOPNN, we combine a fuzzy system and neural networks with the SOPNN. Input variables are partitioned into several subspaces by the fuzzy system or neural network, and these subspaces are utilized as new input variables to the SOPNN architecture. Two types of the advanced SOPNN are obtained by combining not only the fuzzy rules of a fuzzy system with SOPNN but also the nodes in a hidden layer of neural networks with SOPNN into one methodology. The proposed method is applied to the nonlinear system with two inputs, which cannot be identified by conventional SOPNN to show the performance of the advanced SOPNN. The results show that the proposed method is efficient for systems with limited data set and a few input variables and much more accurate than other modeling methods with respect to identification error.     

Comparison of layered cellular manufacturing system design approaches

In this study, a mathematical programming approach is proposed to design a layered cellular manufacturing system in highly fluctuated demand environment. A mathematical model is developed to create dedicated, shared and remainder cells with the objective of minimizing the number of cells. In contrast with classical cellular manufacturing systems, in layered cellular systems, some cells can serve to multiple part families. A five-step hierarchical methodology is employed: (1) formation of part families, (2) calculation of expected cell utilizations and demand coverage probabilities, (3) specification cell types as dedicated, shared, and remainder cells, (4) simulation of proposed layered systems to evaluate their performance with respect to average flowtime and work-in-process inventory, and (5) statistical analysis to find the best layered cellular design among alternatives. It is found that designs with higher number of part families tend to have less number of machines. Similar results are also observed with respect to average flowtime and work-in-process inventory measures. The results are also compared with a heuristic approach from the literature. None of the approaches is dominant with respect to all of the performance measures. Mathematical modeling approach performs better in terms of number of machines for most of the alternative designs. However, heuristic approach yields better average flowtime and work-in-process inventory for most of the designs.     

Designing cellular manufacturing systems under dynamic and uncertain conditions

The paper proposes a fuzzy programming based approach to design a cellular manufacturing system under dynamic and uncertain conditions. The dynamic condition indicates a multi-period planning horizon, in which the product mix and demand in each period can be different. As a result, the best cells designed for one period may not be efficient cells for subsequent periods and some of reconfigurations are required. Uncertain condition implicates to the imprecise nature of the part demand and also the availability of the manufacturing facilities in each period planning. An extended mixed-integer programming model of dynamic cellular manufacturing system, in which some of the coefficients in objective function and constraints are fuzzy quantities, is solved by a developed fuzzy programming based approach. The objective is to determine the optimal cell configuration in each period with maximum satisfaction degree of the fuzzy objective and constraint. To illustrate the behavior of the proposed model and verify the performance of the developed approach, a number of numerical examples are solved and the associated computational results are reported.     

Control-affine neural network approach for nonminimum-phase nonlinear process control

The design of controllers for nonlinear, nonminimum-phase processes is very challenging and remains as one of the more difficult control research problems. Most currently available control algorithms rely implicitly or explicitly upon an inverse of the process. Linear control methods for nonminimum-phase processes are typically based on a decomposition of the process into a minimum-phase and a nonminimum-phase part, and subsequent inversion of the minimum-phase component. A similar scheme for nonlinear systems is still an open problem. In this work, an internal model control strategy employing a minimum-phase model is proposed. The minimum-phase model is first-order, minimum-phase and control-affine but statically equivalent to the original process. Because the model is identified directly from input-output data, a first principles model of the process is not required. The inverse of the process is obtained through analytical inversion of the process model. The proposed control scheme is applied to a van de Vusse reactor and a complex continuous stirred tank bioreactor.     

一类非线性系统的神经网络自适应区间观测器设计

本文研究了一类单输入单输出非线性系统的神经网络自适应区间观测器设计问题. 针对由状态和输入所描述的未知非线性函数的界不可测, 现有的区间观测器方法并未有效地处理系统含有参数不确定性的未知非线性函数. 首先, 本文构造两个径向基函数神经网络来逼近未知非线性部分, 进而分别估计系统状态的上下界; 然后, 选择合适的Lyapunov函数, 采用网络权值校正和网络误差选择机制确保所设计的误差动态系统有界和非负性, 并证明了神经网络自适应区间观测器的稳定性; 最后, 通过仿真实例验证了所提出的神经网络自适应区间观测器的有效性.