You are here

Stochastic modelling and simulation

21 January, 2016 - 14:46

We approached e-courses in domains of biostatistics, stochastic modelling and simulation, using as infrastructure a set of Java class libraries for stochastic modelling and simulation, created and implemented by us (Prodan & Prodan Radu, 2002). The papers (Prodan et al., 1999, 2000) demonstrate the advantages of stochastic models for representation of real world activities and focuses on a Java package, which includes a collection of classes for stochastic modelling and simulation. In order to employ mathematics and statistics to analyze some phenomena processes and activities of the real world, we first construct a stochastic model. Once the theoretical model has been constructed, in theory we are able to determine analytically the answers to a lot of questions related to these phenomena and processes. However, in practice is very difficult to get analytically the answers for many of our questions. This is the reason why we must implement the probabilistic mechanism using a programming language, then to perform a simulation study on a computer. Due to actual spread of fast and inexpensive computational power everywhere in the world, the best approach is to model the real phenomenon as faithfully as possible, and then rely on a simulation study to analyze it. We created and implemented a collection of Java class libraries for stochastic modelling and simulation. The stochastic models constructed accurately represent real world phenomena processes and activities, particularly in medicine, pharmacy and health care. We use these Java libraries as an infrastructure to build intelligent and practical e-learning tools, integrated in an electronic educational environment.

We have considered three levels of simulation. The first level consists of simulating random numbers, as they are the basis of any stochastic simulation study (Figure 3.2). Based on the elements of the first level, we created a second level of simulation applied for distributional models, stochastic processes and Monte Carlo methods. We implemented a hierarchy of Java classes which model the classical distributions and we created a collection of Java class libraries for stochastic modelling and simulation (Prodan & Prodan Rodica, 2001).

media/image2.png
Figure 3.2 The simulation levels 

The first two levels of simulation are the basis for the third level, which is devoted to applications. We used distributional models, stochastic processes and Monte Carlo methods to implement some stochastic modelling applications that accurately represent real world processes, phenomena and activities, particularly in medicine, pharmacy and health care (Prodan & Prodan Rodica, 2001; Gorunescu et al., 2002).