A two-machine flowshop maximum tardiness scheduling problem with a learning effect

The primary objective of this paper is to study a two-machine flowshop scheduling problem with a learning effect where the goal is to find a sequence that minimizes the maximum tardiness. We employ a branch-and-bound method and a simulated annealing (SA) method to search for the optimal solution and a near-optimal solution, respectively. Computational results, using Fisher’s (Math Program 11:229–251 1971) framework, show that the mean and maximum number of nodes for the branch-and-bound algorithm decrease when the learning effect is stronger, the value of the tardiness factor is smaller, or the value of the due date range is larger. In addition, comparisons between the SA method and the earliest due date first (EDD) rule are provided for large-job sizes. Results indicate that the percentage of time that the SA solution outperforms the EDD solution decreases as the job size increases and the learning effect becomes greater. Additionally, the SA solution is never worse than the EDD solution.     

基于遗传算法的作业车间提前/拖期调度问题

针对带有交货期窗口硬约束并对提前/拖期零件进行惩罚的一类作业车间调度问题,设计了一种改进型遗传算法,该算法采用＂逆向后推＂和＂顺向前拉＂相结合的两阶段求解策略。针对部分染色体在解码过程中违反交货期窗口硬约束而产生非法解的问题,采用基于关键路径的染色体修复方法来调整染色体基因序列,以期实现在满足交货期窗口硬约束的同时降低零件拖期成本;在保持第一调度阶段拖期成本不变的基础上,采用基于逆向重调度的目标值修订方法来延迟零件开工时间,以降低在制品流动成本和成品提前库存成本。通过80组调度测试用例的仿真结果表明,该算法在降低调度总成本和拖期成本方面具有一定的优势。     

Dynamic job-shop scheduling with sequence-dependent setup times: simulation modeling and analysis

This paper presents the salient aspects of a simulation-based experimental study of scheduling rules for scheduling a dynamic job shop in which the setup times are sequence-dependent. A discrete event simulation model of the job shop system is developed for the purpose of experimentation. Seven scheduling rules from the literature are incorporated in the simulation model. Five new setup-oriented scheduling rules are proposed and implemented. Simulation experiments were conducted under various experimental conditions characterized by factors such as shop load, setup time ratios, and due date tightness. The results indicate that setup-oriented rules provide better performance than ordinary rules. The difference in performance between these two groups of rules increases with the increase in shop load and setup time ratio. One of the proposed rules performs better for mean flow time and mean tardiness measures.     

A neural network model based predictive control approach: application to a semi-batch reactor

Neural networks can be considered to be new modelling tools in process control and especially in non-linear dynamical systems cases. Their ability to approximate non-linear functions has been very often demonstrated and tested by simulation and experimental studies. In this paper, a predictive control strategy of a semi-batch reactor based on neural network models is proposed. Results of a non-linear control of the reactant temperature of a semi-batch reactor are presented. The process identification is composed of an off-line phase that consists in training the network, and of an on-line phase that corresponds to the neural model adaptation so that it fits any modification of the process dynamics. Experimental results when using this method to control a semi-batch reactor are reported and show the great potential of this strategy in controlling non-linear processes.     

The recycling of wool clothes: an artificial neural network colour classification tool

Recycling of clothes is a straightforward approach for the supply of a coloured raw material which does not involve the cost of the colouring process. A real time and completely automated colour classification tool for woollen clothes to be recycled is proposed. The tool uses the combination of a statistical method, called matrix approach, of a self-organizing feature map (SOFM) and a feed-forward backpropagation artificial neural network (FFBP ANN)-based approach, to correctly classify the clothes by respecting the selection criteria provided by human know-how. The developed tool, which uses an appositely developed workbench with a spectrophotometer, is aware of the way the different coloured clothes to be recycled combine each other to create a new one. The tool has been validated using a set of 5,000 differently coloured clothes to be recycled and the classification error in classifying the clothes is within 5%, i.e., lower than the one resulting from the use of an expert human operator.     

Job shop scheduling based on earliness and tardiness penalties with due dates and deadlines: an enhanced genetic algorithm

This paper studies a job shop scheduling problem with due dates and deadlines in the presence of tardiness and earliness penalties. Due dates are desired completion dates of jobs given by the customer, while deadlines are determined by the manufacturer based on customer due dates. Due dates can be violated at the cost of tardiness, whereas deadlines must be met and cannot be violated. The aforementioned scheduling problem, which is NP-hard, can be formulated with the objective function of minimizing the sum of weighted earliness and weighted tardiness of jobs subject to due dates and deadlines. In order to solve this problem, an enhanced genetic algorithm (EGA) is introduced in this paper. EGA utilizes an operation-based scheme to represent schedules as chromosomes. After the initial population of chromosomes is randomly generated, each chromosome is processed through a three-stage decoder, which first reduces tardiness based on due dates, second ensures deadlines are not violated, and finally reduces earliness based on due dates. After the population size is reached, EGA continues with selection, crossover, and mutation. The proposed algorithm is tested on 180 job shop scheduling problems of varying sizes and its performance is discussed.     

Control chart pattern recognition using an optimized neural network and efficient features

Automatic recognition of abnormal patterns in control charts has seen increasing demands nowadays in manufacturing processes. This study investigates the design of an accurate system for control chart pattern (CCP) recognition from two aspects. First, an efficient system is introduced that includes two main modules: the feature extraction module and the classifier module. The feature extraction module uses the entropies of the wavelet packets. These are applied for the first time in this area. In the classifier module several neural networks, such as the multilayer perceptron and radial basis function, are investigated. Using an experimental study, we choose the best classifier in order to recognize the CCPs. Second, we propose a hybrid heuristic recognition system based on particle swarm optimization to improve the generalization performance of the classifier. The results obtained clearly confirm that further improvements in terms of recognition accuracy can be achieved by the proposed recognition system.     

A hybrid approach to the development of a multilayer neural network for wear and fatigue prediction in metal forming

In this paper an approach to surface damage prediction is proposed for the case of metal forming. The method is mainly based on three fundamental stages: (a) the detection of a feasible physical model which is able to give some important understanding of the phenomenon, although with limited generality; (b) the extensive development of an organized experimental campaign, which is necessary to tune up the developed model; and (c) the organization of an efficient and intelligent way of data collecting. The three aspects of the research work have been integrated by means of a neural network which is trained by using data coming from the real plant, from the standard tribometers, and from the reference numerical model. In this sense, the neural network is indented as hybridized. Predictions are shown to be very close to the experimental data obtained in the production plant. The method is useful for minimizing the number of experiments in the process of materials and treatment selection, and in maintenance.     

Machine fault classification using an ART 2 neural network

Machine monitoring and diagnostics has been considered to be an integral part of the manufacturing process in recent years. It has played an important role in increasing productivity and reducing costs. This paper presents a methodology which is built upon parametric modelling and neural network technology for automatic detection and identification of machine faults. An adaptive resonance theory (ART) neural network architecture is used to identify machine faults from the parameters of a parametric model of the vibration signal. The experimental results indicate that the ART 2 neural network is capable of classifying a fault correctly and rapidly by using the parameters of the parametric model of process signals.     

Real-time control and scheduling of flexible manufacturing systems: An ordinal optimisation based approach

A unique real-time control and scheduling framework for flexible manufacturing systems (FMS) is presented in this paper. The framework enables the adoption of different scheduling policies for short-term intervals when responding to the dynamic changes of the FMS shop floor status. Each time when rescheduling is called for, standard clock (SC) simulation is first employed to evaluate the performance of a set of scheduling policies for a short planning horizon. The ordinal optimisation concept is then used to choose quickly the most desirable scheduling policy. Owing to the use of the standard clock technique and the ordinal optimisation concept, this framework accomplishes a dramatic reduction in the time needed for decision making, the essential requirement for real-time control. It is also found that as the scale of the problem increases, the decision-making time increases linearly rather than exponentially. These two important features indicate that this framework has the potential for being successfully implemented in real FMS settings. Although the framework cannot always guarantee the global best performance, the case study indicates that satisfactory performance results are always achieved by using this framework.     

Multi-objective optimization for turning processes using neural network modeling and dynamic-neighborhood particle swarm optimization

In this paper, we introduce a procedure to formulate and solve optimization problems for multiple and conflicting objectives that may exist in turning processes. Advanced turning processes, such as hard turning, demand the use of advanced tools with specially prepared cutting edges. It is also evident from a large number of experimental works that the tool geometry and selected machining parameters have complex relations with the tool life and the roughness and integrity of the finished surfaces. The non-linear relations between the machining parameters including tool geometry and the performance measure of interest can be obtained by neural networks using experimental data. The neural network models can be used in defining objective functions. In this study, dynamic-neighborhood particle swarm optimization (DN-PSO) methodology is used to handle multi-objective optimization problems existing in turning process planning. The objective is to obtain a group of optimal process parameters for each of three different case studies presented in this paper. The case studies considered in this study are: minimizing surface roughness values and maximizing the productivity, maximizing tool life and material removal rate, and minimizing machining induced stresses on the surface and minimizing surface roughness. The optimum cutting conditions for each case study can be selected from calculated Pareto-optimal fronts by the user according to production planning requirements. The results indicate that the proposed methodology which makes use of dynamic-neighborhood particle swarm approach for solving the multi-objective optimization problems with conflicting objectives is both effective and efficient, and can be utilized in solving complex turning optimization problems and adds intelligence in production planning process.     

Comparative analysis of neural network and regression based condition monitoring approaches for wind turbine fault detection

This paper presents the research results of a comparison of three different model based approaches for wind turbine fault detection in online SCADA data, by applying developed models to five real measured faults and anomalies. The regression based model as the simplest approach to build a normal behavior model is compared to two artificial neural network based approaches, which are a full signal reconstruction and an autoregressive normal behavior model. Based on a real time series containing two generator bearing damages the capabilities of identifying the incipient fault prior to the actual failure are investigated. The period after the first bearing damage is used to develop the three normal behavior models. The developed or trained models are used to investigate how the second damage manifests in the prediction error. Furthermore the full signal reconstruction and the autoregressive approach are applied to further real time series containing gearbox bearing damages and stator temperature anomalies.The comparison revealed all three models being capable of detecting incipient faults. However, they differ in the effort required for model development and the remaining operational time after first indication of damage. The general nonlinear neural network approaches outperform the regression model. The remaining seasonality in the regression model prediction error makes it difficult to detect abnormality and leads to increased alarm levels and thus a shorter remaining operational period. For the bearing damages and the stator anomalies under investigation the full signal reconstruction neural network gave the best fault visibility and thus led to the highest confidence level.     

Optimum tire contour design using systematic stom and neural network

An efficient multi-objective optimization method is presented making use of neural network and a systematic satisficing trade-off method (STOM), in order to simultaneously improve both maneuverability and durability of tire. Objective functions are defined as follows: the sidewall-carcass tension distribution for the former performance while the belt-edge strain energy density for the latter. A back-propagation neural network model approximates the objective functions to reduce the total CPU time required for the sensitivity analysis using finite difference scheme. The satisficing trade-off process between the objective functions showing the remarkably conflicting trends each other is systematically carried out according to our aspiration-level adjustment procedure. The optimization procedure presented is illustrated through the optimum design simulation of a representative automobile tire. The assessment of its numerical merit as well as the optimization results is also presented.     

模糊神经网络在电气设备故障检测与诊断中的应用

电力变压器故障的机理难以应用准确的数学模型加以描述,故障现象与故障原因之间存在着很多不确定因素。本文应用人工网络理论,并利用模糊理论预处理数据,建立了基于模糊神经的变压器故障检测诊断模型。结果表明,该方法对变压器进行故障检测诊断是有效的,同时对其它电气设备的故障诊断也具有参考意义。     

Experimental investigation of an FMS due-date scheduling problem: Evaluation of machine and AGV scheduling rules

Although extensive research has been conducted to solve design and operational problems of automated manufacturing systems, many of the problems still remain unsolved. This article investigates the scheduling problems of flexible manufacturing systems (FMSs). Specifically, the relative performances of machine and automated guided vehicle (AGV) scheduling rules are analyzed against various due-date criteria. First, the relevant literature is briefly reviewed, and then the rules are tested under different experimental conditions by using a simulation model of an FMS. The sensitivity to AGV workload, buffer capacity, and processing-time distribution is also investigated to assess the robustness of the scheduling rules.     

Modeling and optimization of Nd:YAG laser micro-weld process using Taguchi Method and a neural network

The use of a pulsed Nd:YAG laser in the 0.1 mm- thick aluminum alloy lap micro-weld process was optimized. The welding parameters that influence the quality of the pulsed Nd:YAG laser lap micro-weld were evaluated by measuring of the tensile-shear strength. In this work, the Taguchi method was adopted to perform the initial optimization of the pulsed Nd:YAG laser micro-weld process parameters. A neural network with a Levenberg-Marquardt back-propagation (LMBP) algorithm was then adopted to develop the relationships between the welding process parameters and the tensile-shear strength of each weldment. The optimal parameters of the pulsed Nd:YAG laser micro-weld process were determined by simulating parameters using a well-trained back-propagation neural network model. Experimental results illustrate the proposed approach.     

A novel approach for coal seam terrain prediction through information fusion of improved D–S evidence theory and neural network

In order to effectively and accurately forecast the distribution of coal seam terrain, a novel prediction approach through information fusion of improved D–S evidence theory and neural network was proposed. An improved strategy based on confidence level was presented for evidence theory to reduce the conflicts between evidences and enhance the fusion effect. Moreover, BPAs function was constructed reasonably through extracting weights from preliminary prediction values of four neural networks, and the flowchart of proposed approach was designed. Furthermore, a simulation example was provided and some comparisons with other fusion prediction methods were carried out. The simulation example and comparison results indicated that the proposed approach was feasible, high-precision and outperforming others. Finally, an industrial application example of mining automation production was demonstrated to specify the effect of proposed system.     

Concurrent design of manufacturing cell and control functions: A neural network approach

This paper presents a neural network based decision support system (DSS) for use in concurrently determining cell configuration, operation plans, and complexity requirements of cell control functions. Advanced simulators and neural network technology are used in developing the DSS. Simulation experiments were conducted with many possible combinations of design changes to generate training pairs for a neural network. Complexity of cell control functions required by each design option was assessed, based on operational requirements, and was used to train another neural net. Once both neural networks are properly trained, one network can be used to predict the cell design configuration given a set of desirable cell performance measures, while the other network can be used to identify complexity requirements of the cell control functions by using the output provided by the first network as input to the second neural net. An operation-driven cell design methodology was applied to sequentially predict requirements of both cell configuration and cell control functions from the trained neural networks. This innovative new design methodology was illustrated via a successful implementation exercise in acquiring a real automated manufacturing cell at industrial settings. The exercise proves that such a DSS serves well as an effective tool for cell designers and the management in determining appropriate cell configuration and cell control functions at the design stage.     

Design of new sound metric and its application for quantification of an axle gear whine sound by utilizing artificial neural network

The gear whine sound of an axle system is one of the most important sound qualities in a sport utility vehicle (SUV). Previous work has shown that, because of masking effects, it is difficult to evaluate the gear whine sound objectively by using only the A-weighted sound pressure level. In this paper, a new objective evaluation method for this sound was developed by using new sound metrics, which are developed based on the increment of signal to noise ration and the psychoacoustic parameters in the paper, and the artificial neural network (ANN) used for the modeling of the correlation between objective and subjective evaluation. This model developed by using ANN was applied to the objective evaluation of the axle-gear whine sound for real SUVs and the output of the model was compared with subjective evaluation. The results indicate a good correlation of over 90 percent between the subjective and objective evaluations. This paper was recommended for publication in revised form by Associate Editor Yeon June Kang Professor Sang-Kwon Lee received a Ph.D. degree in ISVR (Institute of Sound and Vibration Research) from Southampton University in 1998. He joined Hyundai Motor Research Center in Korea, working with the Automotive Noise and Vibration Control Group from 1985 to 1994. He has been the Professor at the Department of Mechanical Engineering, Inha University, Inchon, Korea, since March 1999. His research interests are the digital signal processing, NVH (noise vibration harahness), condition monitoring, product sound quality design and active control.     

A hybridized artificial neural network and imperialist competitive algorithm optimization approach for prediction of soil compaction in soil bin facility

We were inspired to furnish information concerning the promising applicability of a hybrid approach involving artificial neural networks (ANNs), with manifold network functions, and a meta-heuristic optimization algorithm for prediction of soil compaction indices. The employed network functions were the prevailed feed-forward network and the novel cascade-forward network algorithms to accommodate multivariate inputs of wheel load, tire inflation pressure, number of passage, slippage, and velocity each at three different levels for estimating the study objectives of soil compaction (i.e. penetration resistance and soil sinkage). The experimentations were carried out in a soil bin facility utilizing a single wheel-tester. Each ANN trials was developed merely and then by merging with the recently introduced evolutionary optimization technique of imperialist competitive algorithm (ICA). The results were compared on the basis of a modified performance function (MSEREG) and coefficient of determination (R²). Our results elucidated that hybrid ICA–ANN further succeeded to denote lower modeling error amongst which, cascade-forward network optimized by ICA managed to yield the highest quality solutions.