Distribution planning of magazines

Abstract

The main aim of this thesis is to determine the best values of total printingvolume of magazines and the amounts that are sent to sales points, through forecastingand determination of optimum service level to be provided. The basic tradeoff in this problem is between the cost incurred by the returns of the unsold copies and the lost sales as a result of insufficient number of copies in a single endpoint. The existing literature sees the problem from the inventory planning perspective and triesto optimize the number of copies to be delivered. Based on the costs, which are actually di±cult to estimate and dynamic in nature, they calculate the optimum values. However, due to the practical problems regarding the demand estimation, cost structureand the uncertainty of the future, instead of giving only an optimum delivery amount, the expected values of returns and lost sales, together with their confidence intervals should be provided to the decision makers.The first part of the thesis deals with forecasting the sales endpoint demands forthe future issue. Especially, the handling of the censored demand in case of selloutis analyzed. Also, studies are carried out for finding the optimum parameters of theforecasting method. The two main strategies used for planning are mentioned: top-down and bottom-up. The top-down approach first determines the printing volume anddistributes it to the endpoints. In the second part, three alternative ways of applyingthe top-down approach are discussed. The third part presents the existing bottom-up optimization approach that is valid under the assumption of normal probability demanddistribution. It then deduces formulas for finding return and lost sales that would occur. Then, a simulation based approach is described that finds these performance measures,under any demand distribution assumption. In the end, a computer based illustration of these methods is given.