Editors' Choice
The following articles have been selected by the editors and are available FREE online. Click on the article titles below to read them.
Volume 23, issue 1, January 2012
On the effect of non-optimal forecasting methods on supply chain downstream demand
Mohammad Ali and John Boylan
Supply chains are embarking on strategies to improve visibility by sharing consumer demand data. Investments in formal information sharing mechanisms are at the top of the agenda for many companies. However, recently a steady stream of research, assuming optimal forecasting methods, shows that demand data can be mathematically inferred from the customers’ orders, hence eliminating the need for formal mechanisms to share demand data. The authors term this mathematical inference of demand from customer orders as Downstream Demand Inference (DDI). In this paper, they explore the phenomenon of DDI for two non-optimal forecasting methods: Simple Moving Averages (SMA) and Single Exponential Smoothing (SES). They show that in the case of SMA forecasts, the demand at the downstream link can be inferred. In the case of SES, downstream demand cannot be inferred and thus needs to be shared. The value of sharing demand information, when SES is employed, is quantified. These are most important findings given the popularity of both estimators in real world practices.
Volume 22, issue 3, July 2011
Linear and Non-Linear Filtering in Mathematical Finance: A review
Paresh Date and K. Ponomareva
The optimal estimation of the unobserved states of a system using observed data or measurements is a primary consideration in many areas of the sciences and engineering. Filtering is a method of obtaining state estimates recursively in time by combining model predictions with noisy observations. Such method is valuable in determining if a data series
contains information about latent or hidden variables. Date and Ponomareva's paper surveys the essential elements of time series filtering with emphasis on quantitative finance applications. Recent empirical studies using market data are presented in the context of capturing accurately yield curve dynamics and financial variable's volatility behaviour. Various approaches in the filtering of non-linear time series are also discussed.
Volume 22, issue 2, April 2011
SI on “Mathematical Modelling Developments in Sport”
Dynamic Bayesian forecasting models of football match outcomes with estimation of the evolution variance parameter
Alun Owen
Read here about the state-of-the-art in soccer results forecasting. In particular Owen's model allows team strengths to vary over time and this provides dynamic forecasts of match results. A key parameter of this model is the evolution variance that broadly controls the extent to which team strengths vary over time. Owen shows that predictive performance depends crucially on this parameter and he discusses its estimation in detail.
Volume 22 Number 1 January 2011
Optimal pricing, return and modular design policy for build-to-order (BTO) products in a two parties supply chain system
I. Konstantaras, K. Skouri,S. Papachristos
This article is concerned with build-to-order (BTO) products that are manufactured according to customers’ requirements but may be returned to the manufacturer and subsequently disassembled into components that keep their full initial value. Such a return policy may offer a significant competitive advantage to the manufacturer but its benefits relate explicitly to the degree of modularity associated with the products under concern. This paper explores the linkage between the optimal return policy and modular design; it builds on earlier work to incorporate the selling price of the BTO product as an additional variable. Optimal solutions are developed for the decision variables and a numerical analysis offers insights into the potential utilisation of the proposed approach.
Volume 21, Number 4, October, 2010
A maintenance model with minimal and general repair
Michael Jong Kim and Viliam Makis
Kim and Makis propose a semi-Markov decision framework which permits a general repair policy to be considered. The model overcomes the limitation of minimal repair models but also considers minimal repair as a special case. Of course the use of a semi-Markov model is computationally demanding so they develop an embedded technique for improved computation which is relatively quicker than the standard policy iteration procedure.
Volume 21, Number 3, July 2010
Calculating the accuracy of hierarchical estimation
L.W.G. Strijbosch and J.J.A. Moors
This paper is concerned with the accuracy of hierarchical estimation. Instead of forecasting demand at the individual Stock Keeping Unit (SKU) level, companies often rely upon the estimation of the collective requirements for a group of products. There are various reasons for doing so such as the richness of information that typically becomes available at the aggregate level. This aggregate forecast is then broken down to produce the desired individual demand forecasts. In previous work, the authors of this paper proposed an approach for estimating both the total demand for a number of items and the fraction of this total that an individual item takes; multiplying these two quantities gives the hierarchical estimate for each individual demand. From the joint distribution of the individual demands, in this paper the authors propose a fast and general method for finding the bias and variance of the corresponding hierarchical estimator. The method is compared with the authors’ previous results and a very useful discussion concludes the paper.
Volume 21, Number 1, January 2010
Modelling two-echelon serial inventory systems with perishable items
Fredrik Olsson
Product perishability is of major concern in many industrial sectors including the food industry. Products such as bread, meat, fresh fruit, ready-to-cook vegetables etc are typical examples of stock keeping units that become unusable after a certain due date. Managing multi-echelon perishable inventory systems is a very difficult task since the product lifetime has to be taken into consideration. The inventory investment in such systems can be very high, which motivates the development of more cost-effective replenishment policies. This paper deals with a continuous review, two-echelon serial (two locations) inventory system with perishable items. Transportation times and the lifetime of items are assumed to be constant. An efficient approximate technique is developed for the evaluation of (S – 1, S) policies in such a context of application. The quality of the approximation is assessed through a simulation study. The results reveal that the approximate technique works well in most cases.