A self-organizing neural network approach for the design of cellular manufacturing systems

The crux problem of group technology (GT) is the identification of part families requiring similar manufacturing processes and the rearrangement of machines to minimize the number of parts that visit more than one machine cell. This paper presents an improved method for part family formation, machine cell identification, bottleneck machine detection and the natural cluster generation using a self-organizing neural network. In addition, the generalization ability of the neural network makes it possible to assign the new parts to the existing machine cells without repeating the entire computational process. A computer program is developed to illustrate the effectiveness of this heuristic method by comparing it with the optimal technique for large-scale problems.     

Cell formation using multiple process plans

Cell formation (CF) consists of identifying machine groups and part families. Many CF procedures use a part machine matrix as an input and attempt to obtain a block diagonal form. But perfect block diagonalization of parts and machines is not possible is many cases. In this paper we consider a generalized cellular manufacturing (CM) problem, in which each part can have alternate process plans and each operation can be performed on alternate machines. Under these conditions the CF problem of assigning parts and machines to each manufacturing cell can be considered as a two stage process. The first stage deals with the problem of determining a unique process plan for each part. The second stage determines the part families and machine cells. In this research a model for forming part families and machine cells is presented considering alternate process plans. The objective is to analyze how alternate process plans influence and enhance the CM process giving better flexibility to the designer while designing cells for CM.     

Stochastic dynamic job shops and hierarchical production planning

This paper presents an asymptotic analysis of hierarchical production planning in a general manufacturing system consisting of a network of unreliable machines producing a variety of products. The concept of a dynamic job shop is introduced by interpreting the system as a directed graph, and the structure of the system dynamics is characterized for its use in the asymptotic analysis. The optimal control problem for the system is a state-constrained problem, since the number of parts in any buffer between any two machines must remain nonnegative. A limiting problem is introduced in which the stochastic machine capacities are replaced by corresponding equilibrium mean capacities, as the rate of change in machine states approaches infinity. The value function of the original problem is shown to converge to that of the limiting problem, and the convergence rate is obtained. Furthermore, near-optimal controls for the original problem are constructed from near-optimal controls of the limiting problem, and an error estimate is obtained on the near optimality of the constructed controls     

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.     

Automated Recurrent Neural Network Design to Model the Dynamics of Complex Systems

A general purpose implementation of the tabu search metaheuristic, called Universal Tabu Search, is used to optimally design a locally recurrent neural network architecture. The design of a neural network is a tedious and time consuming trial and error operation that leads to structures whose optimality is not guaranteed. In this paper, the problem of choosing the number of hidden neurons and the number of taps and delays in the FIR and IIR network synapses is formalised as an optimisation problem, whose cost function to be minimised is the network error calculated on a validation data set. The performance of the proposed approach has been tested on the problem of modelling the dynamics of a non-isothermal, continuously stirred tank reactor, in two different operating conditions: when a first order exothermic reaction is occurring; and when two consecutive first order reactions lead to a chaotic behaviour. Comparisons with alternative neural approaches are reported, showing the usefulness of the proposed method.     

Design of parts for cellular manufacturing using neural network-based approach

A neural network approach is applied to the problem of integrating design and manufacturing engineering. The self organising map (SOM) neural network recognizes products and parts which are modeled as boundary representation (B-rep) solids using a modified face complexity code scheme adopted, and forms the necessary feature families. Based on the part features, machines, tools and fixtures are selected. These information are then fed into a four layer feed-forward neural network that provides a designer with the desired features that meet the current manufacturing constraints for design of a new product or part. The proposed methodology does not involve training of the neural networks used and is seen to be a significant potential for application in concurrent engineering where design and manufacturing are integrated.     

Modeling and intellligent chatter control strategies for a lathe machine

In the active chatter control of machine tools, the most effective way to suppress the chatter is to place the actuator as close as possible to the tool tip. However, in practice, it is almost impossible to put the actuator at the same location of the tool tip. Also, in many machines the cutting tools are usually long and may be flexible. Both of these problems pose serious problems in machine chatter control. In order to control the chatter effectively and efficiently, a systematic methodology is proposed in this paper to deal with the modeling and control system design aspects of this challenging problem. Because of the flexibility effect in the tool shaft, conventional active control approaches may not perform in an efficient and effective manner. Here, two advanced control algorithms (LMS adaptive filter and fuzzy CMAC neural network) are proposed to counteract this problem. Experimental results on a lathe machine are also included. Approximately 20 dB reduction in chatter has been achieved.     

神经网络集成与支持向量机在多值分类问题上的比较研究

神经网络集成和支持向量机都是在机器学习领域很流行的方法。集成方法成功地提高了神经网络的稳健性和精度,其中选择性集成方法通过算法选择差异度大的个体,取得了很好的效果。而支持向量机更是克服了神经网络的局部最优,不稳定等缺点,也在多个方面取得了很好的结果。该文着重研究这两种方法在小样本多类数据集上的性能,在四个真实数据集上的结果表明,支持向量机性能要比神经网络集成稍好.     

单通道VLSI阵列容错的神经网络算法

为了解决大面积的VLSI（超大规模集成电路）电路制造过程中因缺陷而造成的成品率低的问题，可以采用降阶和冗余两种VLSI阵列重构方法，这两种方法都属于NP-完全问题。该文是通过冗余修复方法来解决这一问题的。该文基于Hopfield神经网络为模型，将阵列的重构问题转化为矛盾图的最大独立集问题。通过Hopfield神经网络的能量函数方程进行求解，求得合理的补偿通道来完成问题的求解。实验分析表明该方法是简单有效的。     

A Neural Network Model For The Common Due Date Job Scheduling On Unrelated Parallel Machines

This paper presents an approach for scheduling under a common due date on parallel unrelated machine problems based on artificial neural network. The objective is to allocate and sequence the jobs on the machines so that the total cost is minimized. The total cost is the sum of the total earliness and the total tardiness cost. The multilayer Perceptron (MLP) neural network is a suitable model in our study due to the fact that the problem is NP-hard. In our study, neural network has been proven to be effective and robust in generating near optimal solutions to the problem.     

The sustainable cell formation problem: manufacturing cell creation with machine modification costs

An approach for manufacturing cell formation with machine modification is presented. In cell formation it is often important in practice to be able to reassign parts to additional machine types in order to create better cell configurations. This involves extending the set of parts that certain individual machines can process. Such extensions may be cheaper than simply purchasing additional machines. Thus, there is the possibility of machine modification to reduce inter-cell travel. The cost of such modifications must be balanced by the consequent reduction in inter-cell travel cost. The extended machine cell formation problem to be described involves the specification of which individual machines should be modified to enable them to process additional part types, part-machine assignment, and the grouping of individual machines for cell formation. The objective is to minimize the sum of the machine modification costs and the inter-cell travel. We call this the sustainable cell formation problem (SCFP). As far as the authors are aware, there have not been any solution procedures for this important problem reported in the open literature. It is our purpose to fill this gap by presenting a mixed integer programming model of the SCFP. We also propose and analyze greedy and tabu search heuristics for the design of large-scale systems related to the SCFP. Computational experience with the solution procedures indicates that they are likely to be useful additions to the production engineer's toolkit.     

A neural network approach to the classification problem

The task of classifying observations into known groups is a common problem in decision making. A wealth of statistical approaches, commencing with Fisher's linear discriminant function, and including variations to accomodate a variety of modeling assumptions, have been proposed. In addition, nonparametric approaches based on various mathematical programming models have also been proposed as solutions. All of these proposed aolutions have performed well when conditions favorable to the specific model are present. The modeler, therefore, can usually be assured of a good solution to his problem of he chooses a model which fits his situation. In this paper, the performance of a neural network as a classifier is evaluated. It is found that the performance of the neural network is comparable to the best of otheother methods under a wide variety of modeling assumptions. The use of neural networks as classifiers thus relieves the modeler of testing assumptions which would otherwise be critical to the performance of the usual classification techniques.     

Heuristic techniques to optimize neural network architecture in manufacturing applications

Nowadays application of neural networks in the manufacturing field is widely assessed even if this type of problem is typically characterized by an insufficient availability of data for a robust network training. Satisfactory results can be found in the literature, in both forming and machining operations, regarding the use of a neural network as a predictive tool. Nevertheless, the research of the optimal network configuration is still based on trial-and-error approaches, rather than on the application of specific techniques. As a consequence, the best method to determine the optimal neural network configuration is still a lack of knowledge in the literature overview. According to that, a comparative analysis is proposed in this work. More in detail four different approaches have been used to increase the generalization abilities of a neural network. These methods are based, respectively, on the use of genetic algorithms, Taguchi, tabu search and decision trees. The parameters taken into account in this work are the training algorithm, the number of hidden layers, the number of neurons and the activation function of each hidden layer. These techniques have been firstly tested on three different datasets, generated through numerical simulations in the Deform2D environment, in an attempt to map the input–output relationship for an extrusion, a rolling and a shearing process. Subsequently, the same approach has been validated on a fourth dataset derived from the literature review for a complex industrial process to widely generalize and asses the proposed methodology in the whole manufacturing field. Four tests were carried out for each dataset modifying the original data with a random noise with zero mean and standard deviation of one, two and five per cent. The results show that the use of a suitable technique for determining the architecture of a neural network can generate a significant performance improvement compared to a trial-and-error approach.     

Irreducible bin packing and normality in routing open shop

The open shop is a classical scheduling problem known since 1976, which can be described as follows. A number of jobs have to be processed by a given set of machines, each machine should perform an operation on every job, and the processing times of all the operations are given. One has to construct a schedule to perform all the operations to minimize finish time also known as the makespan. The open shop problem is known to be NP-hard for three and more machines, while is polynomially solvable in the case of two machines. We consider the routing open shop problem, being a generalization of both the open shop problem and the metric traveling salesman problem. In this setting, jobs are located at nodes of a transportation network and have to be processed by mobile machines, initially located at a predefined depot. Machines have to process all the jobs and return to the depot to minimize makespan. A feasible schedule is referred to as normal if its makespan coincides with the standard lower bound. We introduce the Irreducible Bin Packing decision problem, use it to describe new sufficient conditions of normality for the two machine problem, and discuss the possibility to extend these results on the problem with three and more machines. To that end, we present two new computer-aided optima localization results.

     

A neural network approach for a robot task sequencing problem

This paper presents a neural network approach with successful implementation for the robot task-sequencing problem. The problem addresses the sequencing of tasks comprising loading and unloading of parts into and from the machines by a material-handling robot. The performance criterion is to minimize a weighted objective of the total robot travel time for a set of tasks and the tardiness of the tasks being sequenced. A three-phased parallel implementation of the neural network algorithm on Thinking Machine's CM-5 parallel computer is also presented which resulted in a dramatic increase in the speed of finding solutions. To evaluate the performance of the neural network approach, a branch-and-bound method and a heuristic procedure have been developed for the problem. The neural network method is shown to give good results and is especially useful for solving large problems on a parallel-computing platform.     

A neural network algorithm for servicing jobs with sequential and parallel machines

We consider the FJ‖C _max problem of optimal servicing with respect to performance for a given set of jobs by sequential and parallel machines. The problem FJ‖C _max is a generalization of the classical J‖C _max problem for the case when the servicing system has not only sequential but also parallel (identical) machines. We propose a two-stage algorithm for a heuristic solution of problem FJ‖C _max On the first stage, we solve the problem J‖C _max, i.e., we assume that the servicing system does not have parallel machines. On the second stage, operations are distributed over parallel machines. On both stages of the algorithm, we use neural network decision making models. The efficiency of a neural network algorithm for the problem J‖C _max and problem FJ‖C _max was evaluated on 20 test examples obtained from 20 known J‖C _max problems by including into the servicing system a random number of copies of sequential machines.     

A neural network approach to freeway network traffic control

The paper investigates the application of a feedforward neural network approach to freeway network control via variable direction recommendations at bifurcation locations. A nonlinear control problem is formulated and solved first by use of computationally expensive nonlinear optimization techniques. A feedforward neural network is then trained by optimally adjusting its weights so as to reproduce the optimal control law for a limited number of traffic scenarios. Generalisation properties of the neural network are investigated and a discussion of advantages and disadvantages compared with alternative control approaches is provided.     

Approaches to modeling train scheduling problems as job-shop problems with blocking constraints

The problem of scheduling a set of trains traveling through a given railway network consisting of single tracks, sidings and stations is considered. For every train a fixed route and travel times, an earliest departure time at the origin and a desired arrival time at the destination are given. A feasible schedule has to be determined which minimizes total tardiness of all trains at their destinations. This train scheduling problem is modeled as a job-shop scheduling problem with blocking constraints, where jobs represent trains and machines constitute tracks or track sections. Four MIP formulations without time-indexed variables are developed based on two different transformation approaches of parallel tracks and two different types of decision variables leading to job-shop scheduling problems with or without routing flexibility. A computational study is made on hard instances with up to 20 jobs and 11 machines to compare the MIP models in terms of total tardiness values, formulation size and computation time.     

A comparison between HMLP and HRBF for attitude control

In this paper the problem of controlling the attitude of a rigid body, such as a Spacecraft, in three-dimensional space is approached by introducing two new control strategies developed in hypercomplex algebra. The proposed approaches are based on two parallel controllers, both derived in quaternion algebra. The first is a feedback controller of the proportional derivative (PD) type, while the second is a feedforward controller, which is implemented either by means of a hypercomplex multilayer perceptron (HMLP) neural network or by means of a hypercomplex radial basis function (HRBF) neural network. Several simulations show the performance of the two approaches. The results are also compared with a classical PD controller and with an adaptive controller, showing the improvements obtained by using neural networks, especially when an external disturbance acts on the rigid body. In particular the HMLP network gave better results when considering trajectories not presented during the learning phase.     

Machine cell formation in generalized group technology

In this paper, we study the configuration of machine cells in the presence of alternative routings for part types. The objective function is the minimization of transportation costs. Limits on cell sizes as well as separation constraints (i.e. machines that are not allowed to be placed in the same cell) and co-location constraints (i.e. machines that must be placed in the same cell) may be imposed. An efficient Tabu Search (TS) algorithm is proposed to solve this problem. Extensive computational experiences with large-size problems show that this method outperforms some existing Simulated Annealing (SA) approaches.