http://repub.eur.nl/res/aut/1244/ List of Publications en http://repub.eur.nl/static-eur/img/logo.png http://repub.eur.nl http://repub.eur.nl/res/pub/15462/ 2004-01-01T00:00:00Z • Response Surface Methodology (RSM) is a tool that was introduced in the early 50´s by Box and Wilson (1951). It is a collection of mathematical and statistical techniques useful for the approximation and optimization of stochastic models. Applications of RSM can be found in e.g. chemical, engineering and clinical sciences. In this paper we are interested in finding the best settings for an automated RSM procedure when there is very little information about the stochastic objective function. We will present a framework of the RSM procedures for finding optimal solutions in the presence of noise. We emphasize the use of both stopping rules and restart procedures. Good stopping rules recognize when no further improvement is being made. Restarts are used to escape from non-optimal regions of the domain. We compare different versions of the RSM algorithms on a number of test functions, including a simulation model for cancer screening. The results show that considerable improvement is possible over the proposed settings in the existing literature. http://repub.eur.nl/res/pub/260/ 2002-12-02T00:00:00Z Direct marketing firms want to transfer their message as efficiently as possible in order to obtain a profitable long-term relationship with individual customers. Much attention has been paid to address selection of existing customers and on identifying new profitable prospects. Less attention has been paid to the optimal frequency of the contacts with customers. We provide a decision support system that helps the direct mailer to determine mailing frequency for active customers. The system observes the mailing pattern of these customers in terms of the well known R(ecency), F(requency) and M(onetary) variables. The underlying model is based on an optimization model for the frequency of direct mailings. The system provides the direct mailer with tools to define preferred response behavior and advises the direct mailer on the mailing strategy that will steer the customers towards this preferred response behavior. http://repub.eur.nl/res/pub/562/ 2002-05-17T00:00:00Z With the advent of one-to-one marketing media, e.g. targeted direct mail or internet marketing, the opportunities to develop targeted marketing campaigns are enhanced in such a way that it is now both organizationally and economically feasible to profitably support a substantially larger number of marketing segments. However, the problem of what segments to distinguish, and what actions to take towards the different segments increases substantially in such an environment. A systematic analytic procedure optimizing both steps would be very welcome.In this study, we present a joint optimization approach addressing two issues: (1) the segmentation of customers into homogeneous groups of customers, (2) determining the optimal policy (i.e., what action to take from a set of available actions) towards each segment. We implement this joint optimization framework in a direct-mail setting for a charitable organization. Many previous studies in this area highlighted the importance of the following variables: R(ecency), F(requency), and M(onetary value). We use these variables to segment customers. In a second step, we determine which marketing policy is optimal using markov decision processes, following similar previous applications. The attractiveness of this stochastic dynamic programming procedure is based on the long-run maximization of expected average profit. Our contribution lies in the combination of both steps into one optimization framework to obtain an optimal allocation of marketing expenditures. Moreover, we control segment stability and policy performance by a bootstrap procedure. Our framework is illustrated by a real-life application. The results show that the proposed model outperforms a CHAID segmentation. http://repub.eur.nl/res/pub/571/ 2002-05-01T00:00:00Z The standard RFM models used by direct marketers include behavioral variables, but ignore the role of marketing communications. In addition, RFM models allow customer responsiveness to vary across different customers, but not across diiferent time periods. Hence, the authors first extend RFM models by incorporating the effects of marketing communications and temporal heterogeneity. Then, using direct-marketing data from a Dutch charity organization, they calibrate the proposed model, and find that it better explains customer behavior because it includes information on both the past behavior and marketing communications. More specifically, they show that direct mail communication builds goodwill, which, in turn, enhances customer's likelihood to buy. However, cumulative exposure to direct mail creates irritation, and erodes goodwill. The two opposite effects induce a cyclic pattern of goodwill formation, which repeats over four quarters. Next, the authors find that, when they control for these communications effects, the standard result - customer's likelihood to buy increases as shopping frequency increases - reverses. That is, in contrast to the extant literature, customers who donate frequently are less likely to donate in the near future. These findings are not only stable over time, but also replicate across two large data sets. Finally, the authors discuss the need for implementing pulsing strategy to mitigate irritation, and the possibility of practicing one-to-one marketing by using information on customer responsiveness, which can be estimated for each customer via the proposed model. http://repub.eur.nl/res/pub/169/ 2002-02-21T00:00:00Z This paper contains a survey of the recent literature on the evaluation of direct marketing campaigns. We give an outline of the various stages included in such a campaign. Next, we review the statistical methods most frequently used and we review the general findings from using these methods. http://repub.eur.nl/res/pub/584/ 2002-02-13T00:00:00Z With the increasing interest in decision support systems and the continuous advance of computer science, revenue management is a discipline which has received a great deal of interest in recent years. Although revenue management has seen many new applications throughout the years, the main focus of research continues to be the airline industry. Ever since Littlewood (1972) first proposed a solution method for the airline revenue management problem, a variety of solution methods have been introduced. In this paper we will give an overview of the solution methods presented throughout the literature. http://repub.eur.nl/res/pub/154/ 2002-01-22T00:00:00Z With the increasing interest in decision support systems and the continuous advance of computer science, revenue management is a discipline which has received a great deal of interest in recent years. Although revenue management has seen many new applications throughout the years, the main focus of research continues to be the airline industry. Ever since Littlewood (1972) first proposed a solution method for the airline revenue management problem, a variety of solution methods have been introduced. In this paper we will give an overview of the solution methods presented throughout the literature. http://repub.eur.nl/res/pub/593/ 2001-12-14T00:00:00Z This paper studies decision rules for accepting reservations for stays in a hotel based on deterministic and stochastic mathematical programming techniques. Booking control strategies are constructed that include ideas for nesting, booking limits and bid prices. We allow for multiple day stays. Instead of optimizing a decision period consisting of a fixed set of target booking days, we simultaneously optimize the complete range of target booking dates that are open for booking at the moment of optimization. This yields a rolling horizon of overlapping decision periods, which will conveniently capture the effects of overlapping stays. http://repub.eur.nl/res/pub/105/ 2001-07-12T00:00:00Z The roster planning process at the Dutch security firm NVD was traditionally carried out by hand. A few years NVD was traditionally carried out by hand. A few years ago, because of changing labor laws in the Netherlands, this became practically impossible. We developed a decision support system which has four main modules. The first one checks given rosters for feasibility with respect to the complicated rules of the current Collective Labor Agreement. A second module generates feasible rosters. The third one evaluates each roster with respect to its cost and ergonomic criteria. Finally, the fourth module uses mathematical programming based methods to select high quality rosters. The DSS has received rave reviews from upper management, security employees as well as the planners, who have gained enormous insight into the planning process. The DSS is currently being implemented and will be operational within the near future. http://repub.eur.nl/res/pub/17112/ 2001-05-31T00:00:00Z The roster planning process at the Dutch security firm NVD was traditionally carried out by hand. A few years ago, because of changing labor laws in The Netherlands, this became practically impossible. We developed a decision support system which has four main modules. The first one checks given rosters for feasibility with respect to the complicated rules of the current Collective Labor Agreement. A second module generates feasible rosters. The third one evaluates each roster with respect its cost and ergonomic criteria. Finally the fourth module uses mathematical programming based methods to select high quality rosters. The DSS has received rave reviews from upper management, security employees as well as the planners, who have gained enormous insight into the planning process. The DSS is currently being implemented and will be operational within the near future. http://repub.eur.nl/res/pub/144/ 2001-01-17T00:00:00Z Abstract This paper studies decision rules for accepting reservations for stays in a hotel based on deterministic and stochastic mathematical programming techniques. Booking control strategies are constructed that include ideas for nesting, booking limits and bid prices. We allow for multiple day stays. Instead of optimizing a decision period consisting of a fixed set of target booking days, we simultaneously optimize the complete range of target booking dates that are open for booking at the moment of optimization. This yields a rolling horizon of overlapping decision periods, which will conveniently capture the effects of overlapping stays. http://repub.eur.nl/res/pub/1664/ 2000-12-15T00:00:00Z Both in business to business and in consumer markets direct mailings are an important means of communication with individual customers. This paper studies the mailing frequency problem that addresses the issue of how often to send a mailing to an individual customer in order to establish a profitable long-term relationship rather than targeting profitable groups of customers at every new mailing instance. The mailing frequency is optimized using long-term objectives but restricts the decisions to the number of mailings to send to the individual over consecutive finite planning periods. A stochastic dynamic programming model that is formulated for this problem is easy to solved for many applications in direct mailing. A particular implementation of the model will provide the direct mailer with controls to stimulate desired response behavior of their customers. The model is calibrated for a large non-profit organization and shows that very large improvements can be achieved by approaching the mailing strategy with the mailing frequency problem, both in the number of mailing to send and in the profits resulting from the responses. http://repub.eur.nl/res/pub/1665/ 2000-12-15T00:00:00Z In developing decision-making models for the evaluation of medical procedures, the model parameters can be estimated by fitting the model to data observed in trial studies. For complex models that are implemented by discrete event simulation (microsimulation) of individual life histories, the Score Function (SF) method can potentially be an appropriate approach for such estimation exercises. We test this approach for a microsimulation model of screening for cancer that is fitted to data from the HIP randomized trial for early detection of breast cancer. Comparison of the parameter values estimated by the SF method and the analytical solution shows that method performs well on this simple model. The precision of the estimated parameter values depends (as expected) on the size of the simulation number of life histories), and on the number of parameters estimated. Using analytical representations for parts of the microsimulation model can increase the precision in the estimation of the remaining parameters. Compared to the Nelder and Mead Simplex method which is often used in stochastic simulation because of its ease of implementation, the SF method is clearly more efficient (ratio computer time: precision of estimates). The additional analytical investment needed to implement the method in an (existing) simulation model may well be worth the effort. http://repub.eur.nl/res/pub/1655/ 2000-05-25T00:00:00Z We consider the Nelder and Mead Simplex Method for the optimization of stochastic simulation models. Existing and new adaptive extensions of the Nelder and Mead simplex method designed to improve the accuracy and consistency of the observed best point are studied. We compare the performance of the extensions on a small microsimulation model, as well as on five test functions. We found that gradually decreasing the noise during an optimization run is the most preferred approach for stochastic objective functions. The amount of computation effort needed for successful optimization is very sensitive to the timing of noise reduction and to the rate of decrease of the noise. Restarting the algorithm during the optimization run, in the sense that the algorithm applies a fresh simplex at certain iterations during an optimization run, has adverse effects in our tests for the microsimulation model and for most test functions. http://repub.eur.nl/res/pub/1647/ 2000-04-12T00:00:00Z We develop a framework for automated optimization of stochastic simulation models using Response Surface Methodology. The framework is especially intended for simulation models where the calculation of the corresponding stochastic response function is very expensive or time-consuming. Response Surface Methodology is frequently used for the optimization of stochastic simulation models in a non-automated fashion. In scientific applications there is a clear need for a standardized algorithm based on Response Surface Methodology. In addition, an automated algorithm is less time-consuming, since there is no need to interfere in the optimization process. In our framework for automated optimization we describe all choices that have to be made in constructing such an algorithm. http://repub.eur.nl/res/pub/1604/ 1999-09-07T00:00:00Z Airline seat inventory control is a very profitable tool in the airline industry. Mathematical programming models provide booking limits or bid-prices for all itineraries and fare classes based on demand forecasts. But the actual revenue generated in the booking process fails to meet expectations. Simple deterministic models based on expected demand generate better revenue than more advanced probabilistic models. Recently suggested models put even more effort into demand forecasting. We will show that the dynamic booking process rather than the demand forecasts needs to be addressed. In particular the nesting strategies applied in booking control will counter-effect the profitability of advanced estimation of booking limits and bid-prices. http://repub.eur.nl/res/pub/1595/ 1999-07-28T00:00:00Z Microsimulation models are increasingly used in the evaluation of cancer screening. Latent parameters of such models can be estimated by optimization of the goodness-of-fit. We compared the efficiency and accuracy of the Response Surface Methodology and the Nelder and Mead Simplex Method for optimization of microsimulation models. To this end, we tested several automated versions of both methods on a small microsimulation model, as well as on a standard set of test functions. With respect to accuracy, Response Surface Methodology performed better in case of optimization of the microsimulation model, whereas the results for the test functions were rather variable. The Nelder and Mead Simplex Method performed more efficiently than Response Surface Methodology, both for the microsimulation model and the test functions. http://repub.eur.nl/res/pub/1522/ 1998-12-31T00:00:00Z In this paper we study a model to estimate the probability that a target group of an advertising campaign is reached by a commercial message a given number of times. This contact frequency distribution is known to be computationally difficult to calculate because of dependence between the viewing probabilities of advertisements. Our model calculates good estimates of contact frequencies in a very short time based on data that is often available. A media planning model that optimizes effective reach as a function of contact frequencies demonstrates the usefulness of the model. Several local search procedures such as taboo search, simulated annealing and genetic algorithms are applied to find a good media schedule. The results show that local search methods are flexible, fast and accurate in finding media schedules for media planning models based on contact frequencies. The contact frequency model is a potentially useful new tool for media planners. http://repub.eur.nl/res/pub/1361/ 1995-01-01T00:00:00Z In multi-echelon distribution systems it is usually assumed that demand is only satisfied from the lowest echelon. In this paper we will consider the case where demand can be satisfied from any level in the system. However, then the problem arises of how to allocate orders from customers to the different locations. A possible way of dealing with this problem consists of using a so-called break quantity rule. This easy implementable rule is to deliver every order with a size exceeding the break quantity from a higher echelon. The use of the break quantity rule now results in a reduction of the demand variability at the retailer and hence less safety stocks need to be held. The concept is studied for a two-echelon distribution system, consisting of one warehouse and one retailer, where the inventory at the retailer is controlled by an order up to level policy, and where at the warehouse there is enough inventory to satisfy all orders from the retailer and the customers. For this system an approximation for the long run average costs as a function of the break quantity is derived, and an algorithm is presented to determine the cost-optimal break quantity. Computational results indicate that the break quantity rule can lead to significant cost reductions. http://repub.eur.nl/res/pub/2238/ 1994-01-01T00:00:00Z