Cihan H. Dagli

Automatic PCB Inspection Algorithms

The importance of the inspection process has been magnified by the requirements of the modern manufacturing environment. In electronics mass-production manufacturing facilities, an attempt is often made to achieve 100% quality assurance of all parts, subassemblies, and finished goods. A variety of approaches for automated visual inspection of printed circuits have been reported over the past two decades. In this survey, algorithms and techniques for the automated inspection of printed circuit boards are examined. A classification tree for these algorithms is presented and the algorithms are grouped according to this classification. This survey concentrates mainly on image analysis and fault detection strategies; these also include state-of-the-art techniques. A summary of the commercial PCB inspection systems is also presented.

Artificial Neural Networks for Intelligent Manufacturing

Preface. Intelligent manufacturing: basic concepts and tools. Intelligent manufacturing system. Intelligent systems architecture design techniques. Basic artificial neural network architectures. Hybrid intelligent systems: tools for decision making in intelligent manufacturing. Neurocomputing for intelligent manufacturing: organization and co-ordination level applications. Conceptual design problems. Machine-part family formation. Process planning. Scheduling. Automated assembly systems. Manufacturing feature identification. Vision based inspection. Performance analysis of artificial neural network methods. Neurocomputing for intelligent manufacturing: execution level applications. Process monitoring and control. Adaptive control in manufacturing. Fuzzy neural control. Neural networks in continuous process diagnostics.

New approaches to nesting rectangular patterns

In this study, two approaches are explored for the solution of the rectangular stock cutting problem: neuro-optimization, which integrates artificial neural networks and optimization methods; and genetic neuro-nesting, which combines artificial neural networks and genetic algorithms. In the first approach, an artificial neural network architecture is used to generate rectangular pattern configurations, to be used by the optimization model, with an acceptable scrap. Rectangular patterns of different sizes are selected as input to the network to generate the location and rotation of each pattern after they are combined. A mathematical programming model is used to determine the nesting of different sizes of rectangular patterns to meet the demand for rectangular blanks for a given planning horizon. The test data used in this study is generated randomly from a specific normal distribution. The average scrap percentage obtained is within acceptable limits. In the second approach, a genetic algorithm is used to generate sequences of the input patterns to be allocated on a finite width with infinite-length material. Each gene represents the sequence in which the patterns are to be allocated using the allocation algorithm developed. The scrap percentage of each allocation is used as an evaluation criterion for each gene for determining the best allocation while considering successive generations. The allocation algorithm uses the sliding method integrated with an artificial neural network based on the adaptive resonance theory (ART1) paradigm to allocate the patterns according to the sequence generated by the genetic algorithm. It slides an incoming pattern next to the allocated ones and keeps all scrap areas produced, which can be utilized in allocating a new pattern through the ART1 network. If there is a possible match with an incoming pattern and one of the scrap areas, the neural network selects the best match area and assigns the pattern. Both approaches gave satisfactory results. The second approach generated nests having packing densities in the range 95–97%. Improvement in packing densities was possible at the expense of excessive computational time. Parallel implementation of this unconventional approach could well bring a quick and satisfactory solution to this classical problem.

Machine-part family formation with the adaptive resonance theory paradigm

Abstract The ARTI neural network paradigm employs a heuristic where new vectors arc compared with group representative vectors for classification. ARTI is adapted for the cell formation problem by reordering input vectors and by using a better representative vector. This is validated with both test cases studied in literaure as well as synthetic matrices. Algoriihmns for effective use of ARTI are proposed. This approach is observed to produce sufficiently accurate results and is therefore promising in both speed and functionality. For the automatic generation of an optimal family formation solution a decision support system can be integrated with ARTI.

An approach to two-dimensional cutting stock problems

One of the resource utilization problems is the location of two-dimensional patterns onto stock sheets with finite dimensions. Stock sheets, in this respect, are depletable resources to be used and the remaining material which is known as the scrap (or trim loss) cannot usually be used later for allocating patterns. Thus, a decrease in the amount of scrap yields a decrease in the raw material cost. In the solution of the problem, usually templates of the patterns to be cut are placed on a stock sheet and then moved until an arrangement is obtained that appears to yield a minimum amount of scrap. Presently, this type of manual solution procedure is used and the algorithms are terminated intuitively when the solution obtained is usually far from the optimum. On the other hand, mathematical programming techniques are generally inadequate for the solution of these problems due to computational burden. Hence, the use of heuristics becomes more appropriate. In this study, a heuristic approach is proposed and th...

Intra-cell manpower transfers and cell loading in labor-intensive manufacturing cells

Labor-intensive manufacturing cells consist of simple machines and equipment that require continuous operator attendance and involvement. Operators are often re-assigned to different machines when a new product is released to the cell. The main reason for this re-assignment is to maximize the output rate of the cell by balancing the flow of products through several machines with varying capacities. In this paper, first a product-sequencing problem with the objective of minimizing the total intra-cell manpower transfers is introduced. A three-phase hierarchical methodology is proposed to solve the problem optimally. Next, manpower transfer matrix values are modified considering the distances traveled among machines. In the second part of the paper, a machine-level-based similarity coefficient that uses the number of machines as a similarity measure is discussed. Later, these coefficients are used during the cell loading process to minimize makespan and also machine and space requirements. Manpower allocation decisions are made along with scheduling decisions that are critical in most labor-intensive manufacturing cells and both approaches are illustrated with an example problem.

Composite stock cutting through simulated annealing

This paper explores the use of Simulated Annealing as an optimization technique for the problem of Composite Material Stock Cutting. The shapes are not constrained to be convex polygons or even regular shapes. However, due to the composite nature of the material, the orientation of the shapes on the stock is restricted. For placements of various shapes, we show how to determine a cost function, annealing parameters and performance.

Stock market prediction using different neural network classification architectures

In recent years, many attempts have been made to predict the behavior of bonds, currencies, stocks, or stock markets. The Standard and Poors 500 Index is modeled using different neural network classification architectures. Most previous experiments used multilayer perceptrons for stock market forecasting. A multilayer perceptron architecture and a probabilistic neural network are used to predict the incline, decline, or steadiness of the index. The results of trading with the advice given by the network is then compared with the maximum possible performance and the performance of the index. Results show that both networks can be trained to perform better than the index, with the probabilistic neural network performing slightly better than the multi layer perceptron.

Manufacturing cell loading rules and algorithms for connected cells

Publisher Summary Cellular Manufacturing (CM) can be defined as the implementation of group technology (GT) principles in a manufacturing environment. Situations that require decisions can be grouped together based on pre-selected, commonly shared criteria, and decision that applies to one situation in the group will apply to all of them in that group. The application of GT to manufacturing is achieved by identifying the items with either similar design or manufacturing characteristics and grouping them into families of like items. The benefits derived from CM include reduced work-in-process inventory and setup time, improved product quality, easier scheduling, better visibility of product schedule status, and quicker feedback of manufacturing deficiencies. The chapter discusses control of manufacturing cells, search priority primary product rule, secondary product rules, primary cell rules, number of feasible products (NFP), product mix (PM), and common cell capacity. The rules described are combined in different ways and 48 possible combinations are created. Twenty-four of the rule combinations are of cell priority type and the remaining 24 are of product priority type.

Genetic neuro-scheduler for job shop scheduling

Abstract This paper describes a hybrid approach between two new techniques, Genetic Algorithms and Artificial Neural Networks, for generating Job Shop Schedules (JSS) in a discrete manufacturing environment based on non-linear multi-criteria objective function. Genetic Algorithm (GA) is used as a search technique for an optimal schedule via a uniform randomly generated population of gene strings which represent alternative feasible schedules. GA propagates this specific gene population through a number of cycles or generations by implementing natural genetic mechanism (i.e. reproduction operator and crossover operator). It is important to design an appropriate format of genes for JSS problems. Specifically, gene strings should have a structure that imposes the most common restrictive constraint; a precedence constraint. The other is an Artificial Neural Network, which uses its highly connected-neuron network to perform as a multi-criteria evaluator. The basic idea is a neural network evaluator which maps a complex set of scheduling criteria (i.e. flowtime, lateness) to evaluate values provided by experienced experts. Once, the network is fully trained, it will be used as an evaluator to access the fitness or performance of those stimulated gene strings. The proposed approach was prototyped and implemented on JSS problems based on different model sizes; namely small, medium, and large. The results are compared to the Shortest Proceesing Time heuristic used extensively in industry.