Detalles de Evento
Impartida por: Paola Paraggio (Università degli Studi di Salerno)
Growth curves with sigmoidal behavior are widely used for data analysis among several fields of application. In general terms, a sigmoidal function is a positive, bounded and differentiable function with positive derivative. Usually, its graph has a characteristic Sshape: it thus shows a slow growth at the beginning and an exponential growth that then slows gradually until it reaches an equilibrium value (usually called carrying capacity). During the years, several sigmoidal curves have been introduced (such as the Gompertz or the Korf curves), but in general with sigmoidal function we refer to the case of the logistic growth. Traditionally, most of the aforementioned curves derive from the solution of ordinary differential equation and, in this sense, they are deterministic. In order to incorporate random influences, the dynamic growth models appeared, among them especially diffusion processes. They are constructed in such a way that their mean function is a certain sigmoidal growth curve. On the other hand, several investigations propose a different stochastic counterpart and they define particular birth-death processes whose mean functions correspond to the proposed growth model. In this talk we will investigate both of these two strategies in the case of the classical logistic growth model. Since there are several real situations in which the maximum level of growth is reached after successive stages, in each of which there is a deceleration followed by an explosion of the exponential type, the sigmoidal curves with more than one inflection point (called multisigmoidal) seem to be interesting in real applications, especially if we consider the logistic curve. The forecasting of oil production has been a problem of great current interest, given its fundamental role in the world’s economy. Oil production is known to be cyclical: in any given system, after it reaches its peak, a decline will begin. With this in mind, the geologist M.K. Hubbert, correctly estimated that oil production in USA would peak around 1970. His prediction was based on the study of a special bell-shaped curve (called Hubbert function) which is the first derivative with respect to the time of the logistic function. Some aspects of this theory has lead researchers to extend/modify the original model. One such aspect is related to the fact that Hubbert provides a forecast with only one peak in oil production, but in other works examples were presented showing that oil production in several countries cannot be represented by a single Hubbert cycle and the peaks can be more than one. Since the peaks of the Hubbert function correspond to the inflection points of the logistic curve, we propose a multisigmoidal logistic growth model useful in application and we define both a birth-death and diffusion process whose mean function is of multisigmoidal logistic type.
17 de diciembre de 2019, 12:00, Seminario de la 1º planta del IEMath-GR