Sinan Kayaligil

Manufacturing lead time estimation using data mining

Abstract We explore use of data mining for lead time estimation in make-to-order manufacturing. The regression tree approach is chosen as the specific data mining method. Training and test data are generated from variations of a job shop simulation model. Starting with a large set of job and shop attributes, a reasonably small subset is selected based on their contribution to estimation performance. Data mining with the selected attributes is compared with linear regression and three other lead time estimation methods from the literature. Empirical results indicate that our data mining approach coupled with the attribute selection scheme outperforms these methods.

Ant Colony Optimization based clustering methodology

A novel ACO based methodology (ACO-C) is proposed for spatial clustering.It works in data sets with no a priori information.It includes solution evaluation, neighborhood construction and data set reduction.It has a multi-objective framework, and yields a set of non-dominated solutions.Experimental results show that ACO-C outperforms other competing approaches. In this work we consider spatial clustering problem with no a priori information. The number of clusters is unknown, and clusters may have arbitrary shapes and density differences. The proposed clustering methodology addresses several challenges of the clustering problem including solution evaluation, neighborhood construction, and data set reduction. In this context, we first introduce two objective functions, namely adjusted compactness and relative separation. Each objective function evaluates the clustering solution with respect to the local characteristics of the neighborhoods. This allows us to measure the quality of a wide range of clustering solutions without a priori information. Next, using the two objective functions we present a novel clustering methodology based on Ant Colony Optimization (ACO-C). ACO-C works in a multi-objective setting and yields a set of non-dominated solutions. ACO-C has two pre-processing steps: neighborhood construction and data set reduction. The former extracts the local characteristics of data points, whereas the latter is used for scalability. We compare the proposed methodology with other clustering approaches. The experimental results indicate that ACO-C outperforms the competing approaches. The multi-objective evaluation mechanism relative to the neighborhoods enhances the extraction of the arbitrary-shaped clusters having density variations.

A SELF-ORGANIZING NEURAL NETWORK APPROACH FOR THE SINGLE AGV ROUTING PROBLEM

Abstract In this research, a special form of Automated Guided Vehicle (AGV) routing problem is investigated. The objective is to find the shortest tour for a single, free-ranging AGV that has to carry out multiple pick and deliver (P&D) requests. This problem is an incidence of the asymmetric traveling salesman problem which is known to be NP-complete. An artificial neural network algorithm based on Kohonens self-organizing feature maps is developed to solve the problem, and several improvements on the basic features of self-organizing maps are proposed. Performance of the algorithm is tested under various parameter settings for different P&D request patterns and problem sizes, and compared with the optimal solution and the nearest neighbor rule. Promising results are obtained in terms of solution quality and computation time.

Conceptualizing manufacturing flexibility: An operational approach and a comparative evaluation

An operational approach to conceptualization and the measurement of manufacturing flexibility is presented. A critical review of selected measures from the literature is provided in the context of the operational approach. A flexibility measure is proposed. A formulation and the features of the proposed measure are discussed. An application of the measure is demonstrated through a hypothetical example. Using the same example, selected flexibility measures from the literature are evaluated. The performance of the measures and the proposed measure are compared.

Virtual reality in requirement analysis for CIM system development suitable for SMEs

This paper presents a methodology, based on Virtual Reality (VR), for representing a manufacturing system in order to help with the requirement analysis (RA) in CIM system development, suitable for SMEs. The methodology can reduce the costs and the time involved at this stage by producing precise and accurate plans, specification requirements, and a design for CIM information systems. These are essentials for small and medium scale manufacturing enterprises. Virtual Reality is computer-based and has better visualization effects for representing manufacturing systems than any other graphical user interface, and this helps users to collect information and decision needs quickly and correctly. A VR-RA tool is designed and developed as a software system to realize the features outlined in each phase of the methodology. A set of rules and a knowledge base is appended to the methodology to remove any inconsistency that could arise between the material and the information flows during the requirement analysis. A novel environment for matching the physical and the information model domains is suggested to delineate the requirements.

An adaptive neighbourhood construction algorithm based on density and connectivity

We introduce a novel neighbourhood construction (NC) algorithm.NC algorithm can work in the data sets with no a priori information.NC is parameter-free, and it yields a unique neighbourhood for each data point.We demonstrate the use of NC on clustering and local outlier detection problems. Display Omitted A neighbourhood is a refined group of data points that are locally similar. It should be defined based on the local relations in a data set. However, selection of neighbourhood parameters is an unsolved problem for the traditional neighbourhood construction algorithms such as k-nearest neighbour and ?-neighbourhood. To address this issue, we introduce a novel neighbourhood construction algorithm. We assume that there is no a priori information about the data set. Different from the neighbourhood definitions in the literature, the proposed approach extracts the density, connectivity and proximity relations among the data points in an adaptive manner, i.e. considering the local characteristics of points in the data set. It is based on one of the proximity graphs, Gabriel graph. The output of the proposed approach is a unique set of neighbours for each data point. The proposed approach has the advantage of being parameter-free. The performance of the neighbourhood construction algorithm is tested on clustering and local outlier detection. The experimental results with various data sets show that, compared to the competing approaches, the proposed approach improves the average accuracy 3-66% in the neighbourhood construction, and 4-70% in the clustering. It can also detect outliers successfully.

A computer-supported methodology for requirements modelling in CIM for small and medium size enterprises: A demonstration in Apparel Industry

This paper discusses the first phase of a computerassisted methodology for implementing CIM within small and medium enterprises. Some of the existing methodologies and tools were furnished by extra items to ease the process for small and medium size manufacturing companies. The methodology aims at using limited expertise, limited staff, and minimising cost. The existing analysis tools i.e. DFDs, ELH, and tools in the GRAI methodology were modified and a new tool was developed PIFR to shed light on the needs and new forms of integration of material and data flows in various parts of the organization. A knowledge base is developed for portability and ease of use. To demonstrate the methodology, a case study was carried out in a firm in apparel industry.

Identifying integration types for requirement analysis in CIM development

This paper presents the second phase of a computer‐ assisted methodology for requirement and design analysis stages of implementing CIM within small and medium‐size enterprises (SME). The main objective is to cover informational and functional analysis during the CIM system life‐cycle. The methodology aims at using limited expertise, limited staff, and expenditure, making it especially suited for introducing CIM in SMEs. The paper emphasises the integration aspects of the methodology as the key factors in the requirement and design analysis stages of the implementation of CIM. Integration has been taken as the need to have some form of operational collaboration between two or more functions. Four types of integration have been identified for reducing the complexity of data communications and for narrowing the gap between analysis and implementation phases. A computer‐supported information requirement analysis tool has been developed for implementing the methodology. A case study was carried out in the apparel industry to test the methodology.

Swarm Intelligence-Based Clustering Algorithms: A Survey

Swarm intelligence (SI) is an artificial intelligence technique that depends on the collective properties emerging from multi-agents in a swarm. In this work, the SI-based algorithms for hard (crisp) clustering are reviewed. They are studied in five groups: particle swarm optimization, ant colony optimization, ant-based sorting, hybrid algorithms, and other SI-based algorithms. Agents are the key elements of the SI-based algorithms, as they determine how the solutions are generated and directly affect the exploration and exploitation capabilities of the search procedure. Hence, a new classification scheme is proposed for the SI-based clustering algorithms according to the agent representation. We elaborate on which representation schemes are used in different algorithm categories. We also examine how the SI-based algorithms, together with the representation schemes, address the challenging characteristics of the clustering problem such as multiple objectives, unknown number of clusters, arbitrary-shaped clusters, data types, constraints, and scalability. The pros and cons of each representation scheme are discussed. Finally, future research directions are suggested.

Lot Splitting Under Learning Effects with Minimal Revenue Requirements and Multiple Lot Types

The lot splitting problem in the presence of learning is addressed. This work is an extension of an approach proposed for splitting in the case of a single item. We address the issue of a minimal revenue requirement from partial deliveries until a predetermined time. This is achieved by imposing a constraint on what is originally an unconstrained optimization problem. When sublots of different items are involved, the optimal splitting decisions have to be combined with the sequencing of the deliveries. Numerical examples are presented to demonstrate the proposed approach.