Formulation of the Learning Problem

The six main objectives of the learning community are (1) to provide each student with both a theoretical, scientific and practical understanding of the nature and range of environmental problems (2) to help students develop critical thinking skills, writing skills, and laboratoiy skills necessary to the recognition, formulation, and assessment of environmental problems and proposed solutions ... 

As we learned in this chapter, the formulation of unsteady distributed problems leads to partial differential equations. The solution of these equations is much more involved than that of ordinary differential equations. Among the techniques available, the analytical and computational methods are most frequently referred to. Exact analytical methods such as separation of variables and transform calculus are beyond the scope of the text. However, the method of complex temperature and the use of charts based on exact analytical solutions, being useful for some practical problems, are respectively discussed in Sections 3.4 and 3.6. Among approximate analytical methods, the integral method, already introduced in Sections 2.4 and 3.1, is further discussed in Section 3.5. The analog solution technique is also briefly treated in Section 3.7. 

The usefulness of the learning machine method for chemical applications has been called in doubt by several authors. Typical predictive abilities of 70 to 95 % may be insufficient for consecutive binary decisions C463. The learning machine may probably be applicable only to simple and very Limited classification problems in chemistry C46, 119, 3183. It seems to be difficult at the moment to formulate a definite opinion about the utility of the learning machine for chemical problems. 

We have put this model into mathematical form. Although we have yet no quantitative predictions, a very general model has been formulated and is described in more detail in Appendix A. We have learned and applied here some lessons from Kilkson s work (17) on interfacial polycondensation although our problem is considerably more difficult, since phase separation occurs during the polymerization at some critical value of a sequence distribution parameter, and not at the start of the reaction. Quantitative results will be presented in a forthcoming pub1ication. 

This Section addresses cases with a continuous performance metric, y. We identify the corresponding problem statements and results, which are compared with conventional formulations and solutions. Then Taguchi loss functions are introduced as quality cost models that allow one to express a quality-related y on a continuous basis. Next we present the learning methodology used to solve the alternative problem statements and uncover a set of final solutions. The section ends with an application case study. 

With this book the reader can expect to learn how to formulate and solve parameter estimation problems, compute the statistical properties of the parameters, perform model adequacy tests, and design experiments for parameter estimation or model discrimination. 

From the above examples we can also learn that the transformation of a physical content from a dimensional into a dimensionless form is automatically accompanied by an essential compression of the statement The set of the dimensionless numbers is smaller than the set of the quantities contained in them, but it describes the problem equally comprehensively. (In the third example the dependency between 5 dimensional parameters is reduced to a dependency between only 2 demensionsless numbers ) This is the proof of the so-called pi theorem (pi after II, the sign used for products), which is formulated in the next Section ... 

Neural networks provide an alternative approach. Neural networks are mathematical constructs that are capable of learning, for themselves, relationships within data. The network makes no assumptions about the functional form of the relationships, but simply tries out a range of models to determine one that will best fit to the existing data that are provided to it. As such, increasingly artificial neural networks (often referred to as ANNs) are used to model complex behaviour in problems like pharmaceuticals formulation and processing. 

ABSTRACT Ttiis communication concerns the basic principles and formulation of a model for a circulating fluidised bed (CFB) biomass gasifier. The main problems are the lot of unknowns concerning the set of kinetic equations needed and the lack of data from commercial CFB biomass gasifiers to check the model. The reacting network to be used in the model is deeply analysed. It is concluded that a self learning or self Urning adaptive model is the best solution... 

---