Modeling Problem - PowerPoint PPT Presentation

1 / 49
About This Presentation
Title:

Modeling Problem

Description:

Emergence in Complex Parts : A Semantic Approach to Modeling Complex Systems Peter N. Kugler, Ph.D. University of Connecticut Aptima Corporation – PowerPoint PPT presentation

Number of Views:20
Avg rating:3.0/5.0
Slides: 50
Provided by: ccsrUiucE
Category:

less

Transcript and Presenter's Notes

Title: Modeling Problem


1
(No Transcript)
2
Modeling Problem How to spell mousetrap with
only three letters?
mou set rap
?
3
How to Approach the Problem?
Step 1 Identify a Set of Symbols (States) that
are used to compose words.
Symbol Set a, b, c, . . . , z
4
Alternative Approach to the problem
Semantic Solution to the Problem
MOUSETRAP
Three Letter Word
CAT
Complex Part(s)
Complex Part(s)
Simple Part(s)
5
Question 2 Is there an Modeling Pipeline that
Connects the two classes of Models T
6
Design Engineering for Simple Parts
Design Engineer
Part functionality is the same in or out of the
system Context free.
Analysis and synthesis are inverse processes.
7
Part Functionality Remains Invariant.
Toaster
Design Engineer
Design Principle Build a stand-alone device
that functions as a Context Free (SIMPLE) Part.
8
Context-Free Transformation (Synthesis
Part?Whole)
Part
System
Space Shuttle Example of Context-Free Design
Engineering
Low Temperature Weather Launch Conditions Emerge
at Launch Time!!
9
Challenger Space Shuttle O-Ring Instability
In Technology this type of transformation results
in Very Dangerous Control Situations.
10
Simple Parts can Reveal their Intrinsic Complexity
Explosion
Elastic Functionality
Transformation of a Simple Part into a Complex
Part.
Emergence of Complex Linkage and new Part
Properties and Functionalities.
11
Simple Versus Complex Parts and Linkages
No Emergence of new Part Functionality.
Simple Linkage
Simple Parts Closed Set of
--Functionalities
--Properties (Observables)
--Meanings
--Properties (Observables)
--Meanings
12
Related Issue Open set of properties
Question Is the open set of
functionalities/properties/meaning of a complex
part in a Super-Positioned State before they
are realized by the complex linkage?
13
Formal System Composed of Simple Parts Numbers
and Operators
Number System
14
Simple (Context Free) Parts
15
The Part Transforms its Functionality in
response to system linkages.
16
COMPLEX (Context-Sensitive) PART Biological
Re-Engineering at the Part level
Undifferentiated Stem Cell
Complex Linkage
Context-Sensitive Transformation of Part
Functionality
--Stem Cell Part In a Nondefinite State of
state of Superimposed Functionality. --Functional
ity is singularized through addition of complex
linkages.
Differentiated Functional Capabilities in
response to different Contexts
17
System with Complex Parts Natural Language
blue
18
Complex (Context Sensitive) Parts
Singularization of Meanings, Functionalities,
Observables
Super-positioned properties/ Functionalities/meani
ngs
19
Returning the the question about Syntactic
Reduction T
Issue When is it appropriate to use a formal
model as a predictor for behavior of a complex
system? Alternatively When does a system with
complex parts (e.g., biology) behavior like a
formal System?
20
Condition for Syntactic
Reduction T
All Complex Parts must have Stationary
Functionality to perform a Syntactic Reduction.
21
SYNTACTIC REDUCTION
Hybrid Equation (NL and FL parts)
1 Bean 3 Bean 4 Beans
Syntactic Model?
1 3 4
Computable if it is only composed of simple
parts
Hybrid Equation (NL and FL parts)
Only and operator parts
Semantic Model
Syntactic Model
22
SUMMARY Syntactic Reduction
All complex linkages must be eliminated so as
to perform a complete syntactic reduction There
can be no complex parts!!!
Semantic Model
Syntactic Model
TSyntactic Reduction
Computable Models
Noncomputable Models
23
Complex/Measurement Linkage
Measured State
M
Part
Measurement Tool
Observer
24
Complex/Measurement Linkage
Measured State
M
Part
Measurement Tool
Observer
25
85 degrees gt 30 degrees
Measured States are composed similar to the
energy/bean equation. They have a simple and
complex part. The simple is computable and the
complex is NOT. The complex part (measurement
tool) can not be embedded in a formal system.
Simple parts
Complex Parts
3.7 inches 3.7 inches
Measurement Tool
Measured State
Unmeasured State
26
(No Transcript)
27
(No Transcript)
28
C
B
A
29
C
B
A
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
Entailment Models of Observability
34
Part III The Dangers of Semantic Instabilities
in Control System Design Nuclear Accident at
3-Mile Island.
World-at-Large
Measurement Instability in a Control System
Perceptual Knowledge Measured States
Measurement Tools
Unmeasured Object/Event
Syntactic and Semantic Knowledge
Measurement Problem in the Control Interface
Design
35
Three Mile Island March 28, 1979 Harrisburg, PA
Failure in System Design and Human Error
36
Three Mile Island Reactor Primary Cooling Loop.
Relief Valve
--Water level in primary cooling loop is measured
indirectly by water level in the pressurizer.
--The measured state (observable) is assumed
to be about the state of the water in the
primary cooling loop.
Pressure
Pressurizer
Water Level
Temperature
Liquid Phase
Primary Cooling
Loop
Pump
Liquid Phase
37
Three Mile Island Reactor Primary Cooling Loop.
Relief Valve
Pressure
Pressurizer
--Relief valve in primary cooling loop
accidentally gets stuck open. --Unknown amount
of water is released. --Core temperature
increases as water level in the pressurizer
decreases.
Water Level
Temperature
Liquid Phase
Pump
Liquid Phase
38
Unknown Amount of Water is Added to Return Water
to Normal Level
Relief Valve
Pressure
Pressurizer
--Water is added to return level to normal range
however, pressure remains low and temperature
critically high. --PUZZLE Pressure is too low,
even though the water level is normal in
pressurizer and temperature remains critically
high!! Control Problem!!
Water Level
Temperature
Liquid Phase
Pump
Liquid Phase
39
Solution 2
Control Engineer ponder whether to add or
subtract water.
Assumption Water-Solid Condition
Decision Problem
Solution 1
Too Little Water
Add Water
First Principle of Reactor Engineering Avoid
Water Solid Condition at all costs, once the
system goes solid you will ALWAYS loose the
reactor--the containment vessel will crack, all
water will be released and melt-down is
inevitable.
CORE
Assumption Core Exposed Condition
Emergency Cooling System
40
Solution 2
Control Engineers ponder whether to add or
subtract water.
CONTROL SOLUTION
Conclusion Too much water Control Order
Subtract Water to avoid going solid.
41
(No Transcript)
42
Engineers hear the banging of STEAM CELLS in
the primary cooling loop and realize that the
actual water condition is too little water
Water Condition is Bi-Phasic?
Relief Valve
Unstable Observable!!
Steam Cells
Pump
Emergency Cooling System
43
Control strategy is suddenly changed Water is
Added.
Relief Valve
Unstable Observable!!
Steam Cells
Pump
Emergency Cooling System
Bi-Phasic State
44
Water returns to a single phase condition and
reactor core temperature returns to safe range.
Primary Coolant Loop
Pump
Emergency Cooling System
Single Phase Liquid State
45
Meaning One Single Phase Water Condition
Same Observable had Two Different Meanings
OBSERVABLE
Reactor Context I
Normal Water Level
--Same Observable, but Different
Contexts!! --Meaning of the Water Level changed
with the Context!! --Polysomy Problem in
Natural Language (Blueness).
Reactor Context II
Meaning Two Bi-Phasic Water Condition
46
3-Mile Island was an example of a Semantically
Induced control instability. Out-of-the-box
control situations usually involve some form of
semantic instability.
Normal Water Level
M Measurement Linkage/ Interface
M
M
Simple Part
Complex Part
New Properties Emerged at the Part Level
47
Biological Versus Technological Design
Context-Free Parts
Context-Sensitive Parts
Biological System Semantic Models
Technological System Syntactic Models
?
Semantic Repository
Syntactic Repository
Man is only man at the surface. Remove the skin
and dissect, and immediately you come to
machinery.
Paul Valery
48
(No Transcript)
49
Emergence through measurement.
Creativity is the search for something that
doesnt exist,
and sometimes finding it.
Write a Comment
User Comments (0)
About PowerShow.com