Academia.eduAcademia.edu

Figure 7 – uploaded by Eduardo Morales

See full PDFdownloadDownload figure
Fig. 11. Subsystems of the Hydrodealkylation of toluene. 2.2.3. Simulation by components Our algorithm simulate individual behaviors of each component in the system using QSIM. This process produces behavioral graphs for each component. Indivi- dual behavior graphs are grouped in subsystems, using the subsystem reduction algorithm.

Figure 11 Subsystems of the Hydrodealkylation of toluene. 2.2.3. Simulation by components Our algorithm simulate individual behaviors of each component in the system using QSIM. This process produces behavioral graphs for each component. Indivi- dual behavior graphs are grouped in subsystems, using the subsystem reduction algorithm.

Related Figures (17)

> Fig. 1. Semantic information associated with each state variable. Each variable (e.g., amount highlighted in the upper half) has information about its meaning (the amount of water in the tank), and information about the substance and related component (e.g., input flow highlighted in the lower half, is 7 associated with a particular tank and with water). The complete specification of a physical component in AGE, requires, besides a qualitative model, the semantic meaning of each state variable and all of its landmark values, as well as its input/output variables in order to An intuitive explanation of a device can be given in terms of causal dependencies of state variables. Given a set of
> Fig. 1. Semantic information associated with each state variable. Each variable (e.g., amount highlighted in the upper half) has information about its meaning (the amount of water in the tank), and information about the substance and related component (e.g., input flow highlighted in the lower half, is 7 associated with a particular tank and with water). The complete specification of a physical component in AGE, requires, besides a qualitative model, the semantic meaning of each state variable and all of its landmark values, as well as its input/output variables in order to An intuitive explanation of a device can be given in terms of causal dependencies of state variables. Given a set of
S.B. Gonzdlez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 1 (1HMH) 1all
S.B. Gonzdlez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 1 (1HMH) 1all
Where V1 = valvel, T= tank, V2 =valve2, Q = flow, K = constant, D = delta (diff.), and P = pressure, where q (Ar), Aro, ! | v1, V2, !), q = qualitative constraint (ejem. add, minus, constant, etc), Ar; are qualitative variables, v; are landmark values. Fig. 4. Causal graph without cycles of the qualitative model of a tank and two valves where @ means derivate. causal graph consider for the explanation and_ the explanation are shown in Fig. 6.
Where V1 = valvel, T= tank, V2 =valve2, Q = flow, K = constant, D = delta (diff.), and P = pressure, where q (Ar), Aro, ! | v1, V2, !), q = qualitative constraint (ejem. add, minus, constant, etc), Ar; are qualitative variables, v; are landmark values. Fig. 4. Causal graph without cycles of the qualitative model of a tank and two valves where @ means derivate. causal graph consider for the explanation and_ the explanation are shown in Fig. 6.
S.B. Gonzalez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence ¥ (
S.B. Gonzalez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence ¥ (
Fig. 7. Three possible behavioral explanations of a tank. The text says that there are tree possible initial states, the first one says roughly: The water level of the tank is empty and increasing, the outflow of the tank is 0 and increasing and the net flow of the tank is between 0 and infinite and decreasing. S.B. Gonzdlez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 1 (1H) Hil
Fig. 7. Three possible behavioral explanations of a tank. The text says that there are tree possible initial states, the first one says roughly: The water level of the tank is empty and increasing, the outflow of the tank is 0 and increasing and the net flow of the tank is between 0 and infinite and decreasing. S.B. Gonzdlez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 1 (1H) Hil
Fig. 8. Explanation of one possible qualitative behavior of the tank. The text roughly says: The water level of the tank is half-full and decreasing, the 1 output flow is between 0 and infinite and decreasing ...then the water level of the tank is between empty and half-full and decreasing ...and then there are 2 possible options: ... 1
Fig. 8. Explanation of one possible qualitative behavior of the tank. The text roughly says: The water level of the tank is half-full and decreasing, the 1 output flow is between 0 and infinite and decreasing ...then the water level of the tank is between empty and half-full and decreasing ...and then there are 2 possible options: ... 1
Some states of the MR system
Some states of the MR system
Segment of the reactor behavior
Segment of the reactor behavior
Fig. 15. Constructing behavior graph of MR system. Table 7
Fig. 15. Constructing behavior graph of MR system. Table 7
3. Fig. 18. MRF system, a cyclic physical device with three components.
3. Fig. 18. MRF system, a cyclic physical device with three components.
S.B. Gonzalez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 0 (HM) 1-iHl
S.B. Gonzalez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 0 (HM) 1-iHl
Z0.025 = 1.96y p; =ni/n, where mm, m2, 3, n4 represent counts for category and n =n, +n .+n3+n4 (Mendehnall and Sincich, 1997). They were presented with 11 systems of different char- acteristics and dimensions to evaluate AGE performance (all the systems presented in this paper plus additional ones). There were also given a questionnaire to assess the utility of AGE and understandability of the different produced explanations. Although AGE has been only assessed by a small group, it received very positive and encouraging comments.
Z0.025 = 1.96y p; =ni/n, where mm, m2, 3, n4 represent counts for category and n =n, +n .+n3+n4 (Mendehnall and Sincich, 1997). They were presented with 11 systems of different char- acteristics and dimensions to evaluate AGE performance (all the systems presented in this paper plus additional ones). There were also given a questionnaire to assess the utility of AGE and understandability of the different produced explanations. Although AGE has been only assessed by a small group, it received very positive and encouraging comments.
S.B. Gonzalez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 0 (1MMH) 1-H useful’’), then the functional explanations (between 64% and 92% consider “very useful’) and finally the causal ~~“ The explanations considered most useful were the 7 behavioral (between 72% and 97% consider “very
S.B. Gonzalez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 0 (1MMH) 1-H useful’’), then the functional explanations (between 64% and 92% consider “very useful’) and finally the causal ~~“ The explanations considered most useful were the 7 behavioral (between 72% and 97% consider “very
Para salir de AGE seleccione del menu la opcidn Safir o bien oprima el boton para cerrar la aplicacién, que se encuentra en la esquina superior de la derecha S.B. Gonzdlez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 1 (1HMH) HMM
Para salir de AGE seleccione del menu la opcidn Safir o bien oprima el boton para cerrar la aplicacién, que se encuentra en la esquina superior de la derecha S.B. Gonzdlez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 1 (1HMH) HMM
S.B. Gonzalez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 1 (1HMH) 1-iHl model and obtained an explanation summary of the behavior. This approach is focused for middle-school education and cannot produce explanations at different abstraction levels, however, it allows active exploration by the student that can help to improve the understanding of different physical phenomena.
S.B. Gonzalez-Brambila, E.F. Morales / Engineering Applications of Artificial Intelligence 1 (1HMH) 1-iHl model and obtained an explanation summary of the behavior. This approach is focused for middle-school education and cannot produce explanations at different abstraction levels, however, it allows active exploration by the student that can help to improve the understanding of different physical phenomena.
Related topics:
EngineeringFunctional AnalysisQualitative ReasoningAge effectCausal AnalysisStructural Decomposition AnalysisQualitative ModelingDependence Graph
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved