Compositional rule of inference

According to the compositional rule of inference, the membership function for the consequence B is obtained over all rules by... 

The result of reasoning Q is specified according to the compositional rule of inference as... 

As a very important special case of the compositional rule of inference may be deduced the generalized modus ponens. Assume we have two propositions. The first proposition (P ) provides information about one of the variables (fact) and the second (P2) involves a relationship between two variables (rule). The aim is to obtain information about the second variable. The reasoning scheme is as follows, whereas X and Y are the two variables with universes of discourse U and V (on the left-hand side the classical scheme is represented for the purpose of comparison) ... 

Add Pa, pb and pc pH first OR 4 second) = 5/36-1-5/36-1-1/36 = 1 1/36. This example shows how elementary events can be grouped together into composite events so as to take advantage of the addition and multiplication rules. Reformulation is powerful because virtually any question can be framed in terms of combinations of and and or operations. With these two rules of probability, y ou can draw inferences about a wide range of probabilistic events. 

The S-N data of die SWNT/epoxy composite is shown in Fig. 13.11. Also included in the figure is the tensile strength data of SWNT, obtained previously. Since the SWNT volume fraction varied from sample to sample, stress on SWNT cannot be inferred directly if die composite stress were used in the S-N plot. Therefore, the maximum cyclic stress of SWNT, calculated using rule-of-mixture, is plotted against the number of cycles to failure of the composite. The Young s modulus of SWNT was estimated as 800 GPa, and the maximum cyclic stress of SWNT was calculated to be between 5.37 and 24 GPa. 

Phase splitting behavior can be inferred from activity coefficients. In general, partial miscibility will not occur whenever the infinite-dilution activity coefficients of the components in solution are less than 7. This is a reliable rule but it depends upon the quality of the activity coefficient data or estimates. If y°° for any one of the components is greater than 7, then partial miscibility may occur at some finite composition. The criterion > 7 often is cited as a general rule indicating a partially... 

Proofs in mill i are given in terms of proof trees that are inference rule composition over judgments. Judgments are pairs of a set of formulas E and a formula

logical consequence of the conjunction of those of E. Inference rules ( -ary) are relations between n + judgments that are noted as follows ... 

In the above series, an important paper of Tyreus and Luyben [5] deals with second-order reactions in recycle systems. Two cases are considered complete one-pass conversion of a component (one recycle), and incomplete conversion of both reactants (two recycles). As general heuristic, they found that fixing the flow in the recycle might prevent snowballing. In the first case, the completely converted component could be fed on flow control, while the recycled component added somewhere in the recycle loop. In the second case, the situation is more complicated. Four reactant feed control alternatives are proposed, but only two workable. This is the case when both reactants are added on level control in recycles (CSl), or when the reactant is added on composition control combined with fixed reactor outlet (CS4). As disadvantage, the production rate can be manipulated only indirectly. Other control structures - with one reactant on flow control the other being on composition (CS2) or level control (CS3) - do not work. The last structure can be made workable if the recycle flow rates are used to infer reactant composition in the reactor. This study reinforces the rule that the flow rate of one stream in a liquid recycle must be fixed in order to prevent snowballing. 

An inference system commonly used to develop fuzzy models is the Mamdani fuzzy inference system. The Mamdani approach was developed in the 1970s and was the first inference method applied to control systems [15]. The Mamdani inference procedure describes the output variables as fuzzy sets. The approach uses max-min composition in which the minimum of the two antecedents is taken for a particular rule and the maximum combination of the rules is determined for aggregating the effects of aU the rules. The effect of the max combiner on the output membership functions is to generate an "envelope" of the fired output membership functions. In order to defuzzify this output set, the centroid (weighted average) of the envelope is found by integrating over the 2-dimensional shape. The defuzzification process of the Mamdani approach is computationally intensive. 

---