Alper Murat

A discrete particle swarm optimization method for feature selection in binary classification problems

This paper investigates the feature subset selection problem for the binary classification problem using logistic regression model. We developed a modified discrete particle swarm optimization (PSO) algorithm for the feature subset selection problem. This approach embodies an adaptive feature selection procedure which dynamically accounts for the relevance and dependence of the features included the feature subset. We compare the proposed methodology with the tabu search and scatter search algorithms using publicly available datasets. The results show that the proposed discrete PSO algorithm is competitive in terms of both classification accuracy and computational performance.

mr2PSO: A maximum relevance minimum redundancy feature selection method based on swarm intelligence for support vector machine classification

This paper presents a hybrid filter-wrapper feature subset selection algorithm based on particle swarm optimization (PSO) for support vector machine (SVM) classification. The filter model is based on the mutual information and is a composite measure of feature relevance and redundancy with respect to the feature subset selected. The wrapper model is a modified discrete PSO algorithm. This hybrid algorithm, called maximum relevance minimum redundancy PSO (mr^2PSO), is novel in the sense that it uses the mutual information available from the filter model to weigh the bit selection probabilities in the discrete PSO. Hence, mr^2PSO uniquely brings together the efficiency of filters and the greater accuracy of wrappers. The proposed algorithm is tested over several well-known benchmarking datasets. The performance of the proposed algorithm is also compared with a recent hybrid filter-wrapper algorithm based on a genetic algorithm and a wrapper algorithm based on PSO. The results show that the mr^2PSO algorithm is competitive in terms of both classification accuracy and computational performance.

A fuzzy-AHP approach to prioritization of CS attributes in target planning for automotive product development

Understanding customer requirements and incorporating them into the conceptual vehicle design is the first step of automotive product development (PD). However, lack of quantitative data and undefined relationships between the attributes makes it difficult to develop a quantitative model for analyzing subjective customer satisfaction (CS) attributes. While researchers and practitioners have accomplished a significant success in terms of developing tool such as quality function deployment (QFD) to capture the voice of customers, and mathematical models for selecting engineering design alternatives, there is limited precedence in terms of prior works on customer satisfaction driven quality improvement target planning and prioritization of customer satisfaction attributes for target planning. This paper presents a fuzzy set theory based analytic hierarchy process (fuzzy-AHP) framework for prioritizing CS attributes in target planning. Furthermore, unlike prior QFD papers, we consider a broad range of strategic and tactical factors for determining the weights. These weights are then incorporated into target planning by identifying the gap in the current CS level. A case example from automotive industry is presented to demonstrate efficacy of the proposed methodology. The framework has been implemented on MS Excel(R) so that the industry can easily adopt it with limited amount of training and at no additional software cost.

Selection of contract suppliers under price and demand uncertainty in a dynamic market

In this paper, we consider a supply contracting problem in which the buyer firm faces non-stationary stochastic price and demand. First, we derive analytical results to compare two pure strategies: (i) periodically purchasing from the spot market; and (ii) signing a long-term contract with a single supplier. The results from the pure strategies show that the selection of suppliers can be complicated by many parameters, and is particularly affected by price uncertainty. We then develop a stochastic dynamic programming model to incorporate mixed strategies, purchasing commitments and contract cancellations. Computational results show that increases in price (demand) uncertainty favor long-term (short-term) suppliers. By examining the two-way interactions of contract factors (price, demand, purchasing bounds, learning and technology effect, salvage values and contract cancellation), both intuitive and non-intuitive managerial insights in outsourcing strategies are derived.

Dynamic routing under recurrent and non-recurrent congestion using real-time ITS information

In just-in-time (JIT) manufacturing environments, on-time delivery is a key performance measure for dispatching and routing of freight vehicles. Growing travel time delays and variability, attributable to increasing congestion in transportation networks, are greatly impacting the efficiency of JIT logistics operations. Recurrent and non-recurrent congestion are the two primary reasons for delivery delay and variability. Over 50% of all travel time delays are attributable to non-recurrent congestion sources such as incidents. Despite its importance, state-of-the-art dynamic routing algorithms assume away the effect of these incidents on travel time. In this study, we propose a stochastic dynamic programming formulation for dynamic routing of vehicles in non-stationary stochastic networks subject to both recurrent and non-recurrent congestion. We also propose alternative models to estimate incident induced delays that can be integrated with dynamic routing algorithms. Proposed dynamic routing models exploit real-time traffic information regarding speeds and incidents from Intelligent Transportation System (ITS) sources to improve delivery performance. Results are very promising when the algorithms are tested in a simulated network of South-East Michigan freeways using historical data from the MITS Center and Traffic.com.

A continuous analysis framework for the solution of location-allocation problems with dense demand

Location-allocation problems arise in several contexts, including supply chain and data mining. In its most common interpretation, the basic problem consists of optimally locating facilities and allocating customers to facilities so as to minimize the total cost. The standard approach to solving location-allocation problems is to model alternative location sites and customers as discrete entities. Many problem instances in practice involve dense demand data and uncertainties about the cost and locations of the potential sites. The use of discrete models is often inappropriate in such cases. This paper presents an alternative methodology where the market demand is modeled as a continuous density function and the resulting formulation is solved by means of calculus techniques. The methodology prioritizes the allocation decisions rather than location decisions, which is the common practice in the location literature. The solution algorithm proposed in this framework is a local search heuristic (steepest-descent algorithm) and is applicable to problems where the allocation decisions are in the form of polygons, e.g., with Euclidean distances. Extensive computational experiments confirm the efficiency of the proposed methodology.

Efficient exact optimization of multi-objective redundancy allocation problems in series-parallel systems

This paper proposes a decomposition-based approach to exactly solve the multi-objective Redundancy Allocation Problem for series-parallel systems. Redundancy allocation problem is a form of reliability optimization and has been the subject of many prior studies. The majority of these earlier studies treat redundancy allocation problem as a single objective problem maximizing the system reliability or minimizing the cost given certain constraints. The few studies that treated redundancy allocation problem as a multi-objective optimization problem relied on meta-heuristic solution approaches. However, meta-heuristic approaches have significant limitations: they do not guarantee that Pareto points are optimal and, more importantly, they may not identify all the Pareto-optimal points. In this paper, we treat redundancy allocation problem as a multi-objective problem, as is typical in practice. We decompose the original problem into several multi-objective sub-problems, efficiently and exactly solve sub-problems, and then systematically combine the solutions. The decomposition-based approach can efficiently generate all the Pareto-optimal solutions for redundancy allocation problems. Experimental results demonstrate the effectiveness and efficiency of the proposed method over meta-heuristic methods on a numerical example taken from the literature.

A Framework for Capturing and Analyzing the Failures Due to System/Component Interactions

To keep up with the speed of globalization and growing customer demands for more technology-oriented products, modern systems are becoming increasingly more complex. This complexity gives rise to unpredictable failure patterns. While there are a number of well-established failure analysis (physics-of-failure) models for individual components, these models do not hold good for complex systems as their failure behaviors may be totally different. Failure analysis of individual components does consider the environmental interactions but is unable to capture the system interaction effects on failure behavior. These models are based on the assumption of independent failure mechanisms. Dependency relationships and interactions of components in a complex system might give rise to some new types of failures that are not considered during the individual failure analysis of that component. This paper presents a general framework for failure modes and effects analysis (FMEA) to capture and analyze component interaction failures. The advantage of the proposed methodology is that it identifies and analyzes the system failure modes due to the interaction between the components. An example is presented to demonstrate the application of the proposed framework for a specific product architecture (PA) that captures interaction failures between different modules. However, the proposed framework is generic and can also be used in other types of PA. Copyright

A multi-dimensional shooting algorithm for the two-facility location-allocation problem with dense demand

We develop an efficient allocation-based solution framework for a class of two-facility location-allocation problems with dense demand data. By formulating the problem as a multi-dimensional boundary value problem, we show that previous results for the discrete demand case can be extended to problems with highly dense demand data. Further, this approach can be generalized to non-convex allocation decisions. This formulation is illustrated for the Euclidean metric case by representing the affine bisector with two points. A specialized multi-dimensional shooting algorithm is presented and illustrated on an example. Comparisons with two alternative methods through a computational study confirm the efficiency of the proposed methodology.

A cost sensitive inpatient bed reservation approach to reduce emergency department boarding times

Emergency departments (ED) in hospitals are experiencing severe crowding and prolonged patient waiting times. A significant contributing factor is boarding delays where admitted patients are held in ED (occupying critical resources) until an inpatient bed is identified and readied in the admit wards. Recent research has suggested that if the hospital admissions of ED patients can be predicted during triage or soon after, then bed requests and preparations can be triggered early on to reduce patient boarding time. We propose a cost sensitive bed reservation policy that recommends optimal bed reservation times for patients. The policy relies on a classifier that estimates the probability that the ED patient will be admitted using the patient information collected and readily available at triage or right after. The policy is cost sensitive in that it accounts for costs associated with patient admission prediction misclassification as well as costs associated with incorrectly selecting the reservation time. Results from testing the proposed bed reservation policy using data from a VA Medical Center are very promising and suggest significant cost saving opportunities and reduced patient boarding times.