Hierarchy in the model

In this section, we further separate the sequencing graph hierarchy into two forms calling hierarchy and control-flow hierarchy. 

A problem arises fw conditionals or loops because their execution delays depend on external signals and events that are not known statically. We further categorize the vertices based on this observation. 

Definition 4.1.2 A vertex has data-independent (fixed) delay the time required to execute its operation is fixed for all input data sequences. Otherwise, the vertex has data-dependent delay. 

A vertex is further classified according to the value of its execution delay for a particular input sequence. If a vertex requires one or more cycles to execute (execution delay 0), then it is called a state vertex. Otherwise, it is called a stateless vertex (execution delay = 0). A graph with only stateless vertices is a stateless gnq h. 

The motivation for defining the two forms of hierarchy above stems from our model of hardware resources. A model in the behavioral description corresponds 

---

On the basis of a reactor model, involving enough structural and functional features of biological and engineering (physical) properties of a biotechnological process, new criteria for a scale-up may evolve. The employment of only one criterion, taken from one arbitrary level of a hierarchy, in the modelling and subsequent scale-up frequently fails, as is apparent from numerous... 

Increase granularity of data analysis to reflect what is sold in the channel. Define the business hierarchy in the model based on profit implications. 

Process operators often have the feeling that a monitored process variable depends on many input variables. The authors have come across a situation where the process operators were under the impression that a polymer quality was affected by 33 process input variables. This is not often the case. The first thing to do is to make a distinction between process input variables and state variables. This also determines the hierarchy of the variables in the model. Figure 19.2 show the hierarchy in the modeling exercise. 

Furthermore, a hierarchy of the potentials is expected, due to the geometry of the fee lattice. As the second, first and fourth can be reached by two first neighbours steps only and the further neighbours require more steps, we observe, especially with the Pt V set, the corresponding hierarchy Vi >> V2jV3,V4 >> V5,V6," - The fact that we observe this fee hierarchy between the V s and their concentration independence gives us confidence in the model and our procedure. 

The solution to the learning problem should provide the flexibility to search for the model in increasingly larger spaces, as the inadequacy of the smaller spaces to approximate well the given data are proved. This immediately calls for a hierarchy in the space of functions. Vapnik (1982) has introduced the notion of structure as an infinite ladder of finitedimensional nested subspaces ... 

In Section 9.1.1 we have introduced a stochastic model for the description of surface reaction systems which takes correlations explicitly into account but neglects the energetic interactions between the adsorbed particles as well as between a particle and a metal surface. We have formulated this by master equations upon the assumption that the systems are of the Markovian type. In the model an infinite set of master equations for the distribution functions of the state of the surface and of pairs of surface sites (and so on) arise. This chain of equations cannot be solved analytically. To handle this problem practically this hierarchy was truncated at a certain level. The resulting equations can be solved numerically exactly in a small region and can be connected to a mean-field solution for large distances from a reference point. 

The next level of detail in the model hierarchy of Fig. 6.2 is the so-called dumped rate models" (third from the bottom). They are characterized by a second parameter describing rate limitations apart from axial dispersion. This second parameter subdivides the models into those where either mass transport or kinetic terms are rate limiting. No concentration distribution inside the particles is considered and, formally, the diffusion coefficients inside the adsorbent are assumed to be infinite. 

This means there is a 95% probability that the true value of the parameter Po between -16.044 and 13.644. Since these two limiting values have opposite signs, and since no value in a confidence interval is more probable than any other, the true value of Po could be zero. In other words, the value bo = -1.200 is not statistically significant and there is not enough evidence for maintaining the Po term in the model. AU the same, it is customarily kept in the equation, for the sake of mathematical hierarchy. 

To account for effects of temperature gradients in the rod, we must move to the next level in the model hierarchy, which is to say that a differential volume must be considered. 

Levels 1 and 2 solutions have one assumption in common The rod temperature below the solvent surface was taken to be uniform. The validity of this modelling assumption will not be known until we move up one more level in the model hierarchy. 

The obvious question arises When is a model of a process good enough This is not a trivial question, and it can only be answered fiiUy when the detailed economics of design and practicality are taken into account. Here, we have simply illustrated the hierarchy of one simple process, and how to find the limits of validity of each more complicated model in the hierarchy. In the final analysis, the user must decide when tractability is more important than precision. 

Table 19.1 shows the experimental design and results of CCD of response surface methodology. The factors levels are 3.5, 5, and 6.5 for pH, 20, 40, and 60 g/1 for L, 0, 100, and 200 mg/1 for NTC, and 0.03, 0.3, and 0.6 g/1 for Y. In the last column, the obtained RF values are shown. The effects of the parameters on the RF were calculated and the parameters which showed P-values less than 0.05 were taken into account in the model the other parameters were actually undistinguishable from noise. The significant terms, as shown in Table 19.2, are pH, Y, NTC, pH, and pH-L. Then, by eliminating the other terms from the model (except L to support hierarchy as requested from the methodology (Box et al., 1978)), Eq. (2) was obtained as a function of the significant factors. 

To define a hierarchy, under a risk context, a model of the consequences is needed. According to Aven (2012), the most appropriate definition of risk involves defining the consequences and uncertainties. The uncertainties come from different sources in the model. In this context, under Decision Theory, the uncertainties considered are those (Garcez Almeida, 2014) t) arising from the occurrence of a hazard scenario (because it is not known what hazard scenario actually will occur), ( ) arising from the consequence of a hazard scenario that has occurred (because there are probabilistic mechanisms that interfere in the magnitude of the consequences). 

During the system analysis one focuses on the system itself, in order to investigate which physical-chemical phenomena take place and are relevant with respect to the modeling goal. One method that could be used is to find key mechanisms and key components, for example evaporation (phase equilibrium), mass diffusion or transport, heat convection, conduction or radiation, and liquid flow. The key variables could be, amongst others, temperature, pressure and concentration at a certain location. Also in this case, one should take into account the required level of detail, the hierarchy within the model and the required accuracy. 

---