McLeod Institute of Simulation Sciences
Academic Certificate

Engineering Degree (five years plus thesis) given by the

University of Genoa

To:

Enrico Bocca

Supervisor:

Agostino Bruzzone

No:

2

Thesis Title

Advanced M&S for innovative technologies applied in Retail sector

Summary
The researcher presents the use frequency domain modeling methods for optimizing the inventory management in retail supply chain, in order to identify periodic component influence and to support demand forecasts within a very stochastic framework. The quantitative analysis as a means of optimizing goods management is a time-tested tool used quite extensively in the logistics area of all companies. In the retail business sector in particular, optimized product management is an important objective since it generates benefits in terms of cost (reduction in locked-up capital), image (optimized quality of the service to the customer) and logistics (problems related to product flows and storage). The logistics aspect is especially important considering the orographic features of Liguria with its severely restricted product storage areas. Thus, this aspect highlights the problems inherent to managing the large quantities of products needed to guarantee service quality in limited spaces. Such approach has been developed to evaluate the efficiency of various predictive algorithms to be applied in order to analyze the demand for food products, from the point of sales to the central warehouse. It focused on 3rd Degree Exponential Smoothing, how to determine the seasonally behavior in relation to the demand for each product. The techniques implemented to highlight this characteristic are broadly applied in the field of signal processing (i.e. audio signals). In fact, if we graphically represent demand of the sales network supplied by the central warehouse versus time, the parameter being studied is comparable to a discrete signal. By implementing the Fourier analysis, in this case applying the Fourier transform for discrete signals (DFT Discrete Fourier Transform), it is possible to divide the signal into a series of components with modulus, frequency and phase. In this way, once the signal has been broken down into its (harmonic) components, and by observing its modulus, it can be identified the (predominant) frequency or frequencies. Applying the DFT to the discrete signal generates the set of components into which the signal can be divided. The result obtained was checked by reversing the anti-transform operation.
Back to MISS Certificates