Haldun Aytug

Executing production schedules in the face of uncertainties: A review and some future directions

We review the literature on executing production schedules in the presence of unforeseen disruptions on the shop floor. We discuss a number of issues related to problem formulation, and discuss the functions of the production schedule in the organization and provide a taxonomy of the different types of uncertainty faced by scheduling algorithms. We then review previous research relative to these issues, and suggest a number of directions for future work in this area. � 2003 Elsevier B.V. All rights reserved.

Use of genetic algorithms to solve production and operations management problems: A review

Operations managers and scholars in their search for fast and good solutions to real-world problems have applied genetic algorithms to many problems. While genetic algorithms are promising tools for problem solving, future research will benefit from a review of the problems that have been solved and the designs of the genetic algorithms used to solve them. This paper provides a review of the use of genetic algorithms to solve operations problems. Reviewed papers are classified according to the problems they solve. The basic design of each genetic algorithm is described, the shortcomings of the current research are discussed and directions for future research are suggested.

Rapid Modeling and Discovery of Priority Dispatching Rules: An Autonomous Learning Approach

Priority-dispatching rules have been studied for many decades, and they form the backbone of much industrial scheduling practice. Developing new dispatching rules for a given environment, however, is usually a tedious process involving implementing different rules in a simulation model of the facility under study and evaluating the rule through extensive simulation experiments. In this research, an innovative approach is presented, which is capable of automatically discovering effective dispatching rules. This is a significant step beyond current applications of artificial intelligence to production scheduling, which are mainly based on learning to select a given rule from among a number of candidates rather than identifying new and potentially more effective rules. The proposed approach is evaluated in a variety of single machine environments, and discovers rules that are competitive with those in the literature, which are the results of decades of research.

A review of machine learning in scheduling

This paper has two primary purposes: to motivate the need for machine learning in scheduling systems and to survey work on machine learning in scheduling. In order to motivate the need for machine learning in scheduling, we briefly motivate the need for systems employing artificial intelligence methods for scheduling. This leads to a need for incorporating adaptive methods-learning. >

Stopping Criteria for Finite Length Genetic Algorithms

Considerable empirical results have been reported on the computational performance of genetic algorithms but little has been studied on their convergence behavior or on stopping criteria. In this paper we derive bounds on the number of iterations required to achieve a level of confidence to guarantee that a genetic algorithm has seen all populations and, hence, an optimal solution.

New stopping criterion for genetic algorithms

Abstract Genetic Algorithms have been successfully applied in a wide variety of problems. Although widely used, there are few theoretical guidelines for determining when to terminate the search. One result by Aytug and Koehler provides a loose bound on the number of GA generations needed to see all populations (and hence, an optimal solution) with a specified probability. In this paper we derive a tighter bound. This new bound is on the number of iterations required to achieve a level of confidence to guarantee that a Genetic Algorithm has seen all strings (and, hence, an optimal solution).

Detecting Management Fraud in Public Companies

This paper provides a methodology for detecting management fraud using basic financial data. The methodology is based on support vector machines. An important aspect therein is a kernel that increases the power of the learning machine by allowing an implicit and generally nonlinear mapping of points, usually into a higher dimensional feature space. A kernel specific to the domain of finance is developed. This financial kernel constructs features shown in prior research to be helpful in detecting management fraud. A large empirical data set was collected, which included quantitative financial attributes for fraudulent and nonfraudulent public companies. Support vector machines using the financial kernel correctly labeled 80% of the fraudulent cases and 90.6% of the nonfraudulent cases on a holdout set. Furthermore, we replicate other leading fraud research studies using our data and find that our method has the highest accuracy on fraudulent cases and competitive accuracy on nonfraudulent cases. The results validate the financial kernel together with support vector machines as a useful method for discriminating between fraudulent and nonfraudulent companies using only publicly available quantitative financial attributes. The results also show that the methodology has predictive value because, using only historical data, it was able to distinguish fraudulent from nonfraudulent companies in subsequent years.

Making words work: Using financial text as a predictor of financial events

We develop a methodology for automatically analyzing text to aid in discriminating firms that encounter catastrophic financial events. The dictionaries we create from Management Discussion and Analysis Sections (MD&A) of 10-Ks discriminate fraudulent from non-fraudulent firms 75% of the time and bankrupt from nonbankrupt firms 80% of the time. Our results compare favorably with quantitative prediction methods. We further test for complementarities by merging quantitative data with text data. We achieve our best prediction results for both bankruptcy (83.87%) and fraud (81.97%) with the combined data, showing that that the text of the MD&A complements the quantitative financial information.

Solving large-scale maximum expected covering location problems by genetic algorithms: A comparative study

Abstract This paper compares the performance of genetic algorithms (GAs) on large-scale maximum expected coverage problems to other heuristic approaches. We focus our attention on a particular formulation with a nonlinear objective function to be optimized over a convex set. The solutions obtained by the best genetic algorithm are compared to Daskins heuristic and the optimal or best solutions obtained by solving the corresponding integer linear programming (ILP) problems. We show that at least one of the GAs yields optimal or near-optimal solutions in a reasonable amount of time.

Accurate estimation of expected coverage: revisited

Abstract As noted in several studies (Batta et al., Transp. Sci. 23 (1989) 277), (Burwell et al., Comput. Opns. Res. 20 (1993) 113), (Daskin, Network and Discrete Location, Wiley, New York, 1995), (Marianov and ReVelle, Eur. J. Opns. Res. 93 (1996) 110), (Saydam et al., Socio-Econ. Plann. Sci. 28(2) (1994) 113), the accurate estimation of expected coverage is an important and open issue. Although the maximum expected coverage model is empirically shown to prescribe a robust set of “optimal” locations, earlier findings suggest that it could also over or underestimate the coverage by a significant margin. In this study, we present a genetic algorithm (GA) that combines the expected coverage approach with the hypercube model (Jarvis, Mgmt. Sci. 31 (1985) 235), (Larson, Comput. Opns. Res. 1 (1974) 67), (Larson, Opns. Res. 23 (1975) 845) to solve the maximum expected coverage location problem with increased accuracy and realism. Our findings suggest that the GA provides at least as good solutions 94% of the time making it a viable alternative to the two-step procedures stipulated earlier.