2008 IEEE International Conference on Technologies for Homeland Security A Modeling Framework for Evaluating Effectiveness of Smart-Infrastructure Crises Management Systems Tridib Mukherjee and Sandeep K. S. Gupta Impact Lab (http:// impact.asu.edu ) School of Computing & Informatics Arizona State University [email protected]
29
Embed
2008 IEEE International Conference on Technologies for Homeland Security A Modeling Framework for Evaluating Effectiveness of Smart-Infrastructure Crises.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
2008 IEEE International Conference on Technologies for Homeland Security
A Modeling Framework for Evaluating Effectiveness of
2008 IEEE International Conference on Technologies for Homeland Security
Outline
• Motivation.
• Smart-Infrastructure Crises Management.
• Criticality Response Modeling (CRM) framework to evaluate crises response for smart-infrastructure.
• Application of CRM to fire emergencies in offshore Oil & Gas Production Platforms (OGPP).
• Simulation based verification of the framework.
• Conclusions & Future Work.
2008 IEEE International Conference on Technologies for Homeland Security
Goals of Homeland Security
• Department of Homeland Security (DHS) missions include– Prevention of terrorist attacks within the US.
– Reduction of vulnerability to terrorism.
– Minimizing the damage from potential attacks and natural disasters.
– In summary: be prepared for potential national crises and planning proper responses.
• DHS combines 22 federal agencies into four policy directorates– Border and Transportation Security.
– Emergency Preparedness and Response.
– Information Analysis and Infrastructure Protection.
– Science and Technology.
2008 IEEE International Conference on Technologies for Homeland Security
Importance of crises response and preparedness to DHS
• In 2004, over $4 billion of Homeland Security Grants allocated for assistance to the first responders.
• In 2005, $7.4 billion fund budgeted for Emergency Preparedness and Response (around 20% of the total budget).– over $3.5 billion (50%) budgeted for assistance to first responders.
• Since March 1, 2003, approximately $8 billion awarded to state, tribal and local governments to prevent, prepare for, respond to and recover from acts of terrorism and all hazards.
2008 IEEE International Conference on Technologies for Homeland Security
What are Crises?
Massive (cascading) catastrophic events leading to loss of lives/property
Evaluate the Q-value of Criticality Response Process
CRM Framework
Evaluate Effectiveness of Response Process
Criticality Response Modeling (CRM) Framework
Lea
rnin
g
2008 IEEE International Conference on Technologies for Homeland Security
Application of CRM
2008 IEEE International Conference on Technologies for Homeland Security
Fire Emergencies in offshore Oil & Gas Production Platforms (OGPP) – example process flow*
* D. G. DiMattia, F. I. Khan, and P. R. Amyotte, “Determination of human error probabilities for offshore platform musters,” Journal of Loss Prevention in the Process Industries, vol. 18, pp. 488–501, 2005.
S tar t
F ir e & Ex p lo s io nAlar m ?
N o
I m m in en tD an g er ?
Yes
R etu r n p r o c es s eq u ip m en tto s af e s ta te
M ak e w o r k p lac e s af e
N o
Ev a lu a te p o te n t ia l e v a c u a t io np a th s a n d c h o o s e ro u te (o r
a lt e rn a te ro u te )
M o v e a lo n g ev ac u atio nr o u te
L is ten & f o llo w P Aan n o u n c em en ts
As s is tan c er eq u ir ed f o r
o th er s ?
Yes
Yes
E v ac u atio np ath n o tten ab le?
N o
Yes
R eg is te r a t tem p o r ar yr efu g e
N o
P r o v id e f eed b ac k
2008 IEEE International Conference on Technologies for Homeland Security
S tar t
F ir e & Ex p lo s io nAlar m ?
N o
I m m in en tD an g er ?
Yes
R etu r n p r o c es s eq u ip m en tto s af e s ta te
M ak e w o r k p lac e s af e
N o
Ev a lu a te p o te n t ia l e v a c u a t io np a th s a n d c h o o s e ro u te (o r
a lt e rn a te ro u te )
M o v e a lo n g ev ac u atio nr o u te
L is ten & f o llo w P Aan n o u n c em en ts
As s is tan c er eq u ir ed f o r
o th er s ?
Yes
Yes
E v ac u atio np ath n o tten ab le?
N o
Yes
R eg is te r a t tem p o r ar yr efu g e
N o
P r o v id e f eed b ac k
CRM for fire emergencies in OGPP – Identify Criticalities
criticality 1 (c1)
criticality 2 (c2)
criticality 3 (c3)
criticality 4 (c4)
Identify the decision boxes of the process flow as criticalities.
2008 IEEE International Conference on Technologies for Homeland Security
CRM for fire emergencies in OGPP – Identify Response Actions
Identify the appropriate decision branches of the process flow as response actions.
S tar t
F ir e & Ex p lo s io nAlar m ?
N o
I m m in en tD an g er ?
Yes
R etu r n p r o c es s eq u ip m en tto s af e s ta te
M ak e w o r k p lac e s af e
N o
Ev a lu a te p o te n t ia l e v a c u a t io np a th s a n d c h o o s e ro u te (o r
a lt e rn a te ro u te )
M o v e a lo n g ev ac u atio nr o u te
L is ten & f o llo w P Aan n o u n c em en ts
As s is tan c er eq u ir ed f o r
o th er s ?
Yes
Yes
E v ac u atio np ath n o tten ab le?
N o
Yes
R eg is te r a t tem p o r ar yr efu g e
N o
P r o v id e f eed b ac k
Response to c1
c1
c2c3
c4
Response to c3, c4
Response to c2
2008 IEEE International Conference on Technologies for Homeland Security
CRM for fire emergencies in OGPP – Identify States and Determine Window-of-opportunity
Fire Alarm
Fire Alarm &Imminent Danger
Fire Alarm &Non-tenable Path
Fire Alarm &Assistance Required
Fire Alarm &Non-tenable Path &
Assistance Required
Fire Alarm &Imminent Danger &
Assistance Required
Fire Alarm &Assistance Required &
Non-tenable Path
Criticalities
1. c1 – Fire Alarm.
2. c2 – Imminent danger e.g. health hazards.
3. c3 – Assistance required to others e.g. trapped personnel.
4. c4 – Evacuation path not tenable.
Window-of-opportunity survival time under
asphyxiation.
2008 IEEE International Conference on Technologies for Homeland Security
CRM for fire emergencies in OGPP – Determine State Transition Probabilities
Fire Alarm
Fire Alarm &Imminent Danger
Fire Alarm &Non-tenable Path
Fire Alarm &Assistance Required
Fire Alarm &Non-tenable Path &
Assistance Required
Fire Alarm &Imminent Danger &
Assistance Required
Fire Alarm &Assistance Required &
Non-tenable Path
State transition probabilities derived from established probability distribution in [1].
[1] D. G. DiMattia, F. I. Khan, and P. R. Amyotte, “Determination of human error probabilities for offshore platform musters,” Journal of Loss Prevention in the Process Industries, vol. 18, pp. 488–501, 2005.
0.1634
0.284877
0.40365
0.1755
0.1634
0.2965 0.48970.5862
0.5717
0.5717 0.2649
0.2649
0.481
0.1977 0.1892
0. 2094
0.41861 0.3348
0.4138
2008 IEEE International Conference on Technologies for Homeland Security
• Response Action Selection Policies– Greedy – response actions corresponding to ML with maximum
probability• Oblivious of subsequent criticalities.
– Mitigative Action based Criticality Management (MACM) – response actions corresponding to MLs with maximum Q-values
• Not oblivious of subsequent criticalities.
• Simulation Goal– Compare different response action selection policies.– Evaluate impact of timing factors to manageability of criticality
– Verifies applicability of Q-value as manageability metric.
Simulation Study
2008 IEEE International Conference on Technologies for Homeland Security
Greedy and MACM action selection Comparison
Low manageability for Greedy response action selection
(sec)
(Q-v
alu
e)
(MACM)
(MACM)
Zero manageability for high detection delay
Low manageability for increase in number of simultaneous criticalities
2008 IEEE International Conference on Technologies for Homeland Security
Effect of Actuation and Detection Delay for two simultaneous criticalities
Low manageability for high action time
(sec)
(sec)
(Q-v
alu
e)
Low manageability for high action time
2008 IEEE International Conference on Technologies for Homeland Security
Effect of Actuation and Detection Delay for three simultaneous criticalities
Low manageability for increase in number of simultaneous criticalities
(sec)(sec)
(Q-v
alu
e)
2008 IEEE International Conference on Technologies for Homeland Security
Conclusions
• CRM framework developed for evaluating effectiveness of crises response processes.
• CRM applied to real crisis situation – fire emergencies in Oil & Gas Production Platforms.
• CRM enables – Q-value based quantitative evaluation of crises response.– automated learning from the outcome.– steeper learning curve – improved preparedness for crises
response.
2008 IEEE International Conference on Technologies for Homeland Security
Future Work
• Q-value calculation computationally expensive– good metric for evaluation.– bad for on-line planning.
• Probabilistic planning to select response actions based on the stochastic model.– determine optimal response selection policy.– computation complexity within temporal requirements.
• Develop simulation tools and visualization of the planned actions and their effects – for use by the disaster manager.
2008 IEEE International Conference on Technologies for Homeland Security
Questions ??
Impact Lab (http://impact.asu.edu)
Creating Humane Technologies for Ever-Changing World
2008 IEEE International Conference on Technologies for Homeland Security
Additional Slides
2008 IEEE International Conference on Technologies for Homeland Security
Effectiveness Evaluation for the Response Actions
• Generally in terms of cumbersome documents– Reports / recommendations– Qualitative & subjective– Inadequate for smart-infrastructure
• Requires quantitative evaluation• Objective comparison between different response actions for steeper learning curve• Evaluate impact of different parameters to the effectiveness of criticality response
• Quantitative Evaluation– What are the evaluation criteria & metrics?
• Theoretical Foundation Established in our previous work – crises characterized as criticalities.– How to perform evaluation for any crises response process?
• Research Goal: Develop generic evaluation framework for crisis response.
• Contributions– Criticality Response Modeling (CRM) Framework– Application of CRM for fire emergencies in offshore Oil & Gas Production Platforms
(OGPP)– Simulation based evaluation of CRM over OGPP
2008 IEEE International Conference on Technologies for Homeland Security
Manageability as Q-value
n
NORMAL STATE
i
x
px,i
• Manageability from any arbitrary critical state x
– i an immediate upstream state.
Qx,i,n = px,iPi,n if W met = 0 if W NOT met
Pi,n = 1 if i = n = (1 - pi,j ) pi,kPk,n + pi,jPj,n if i n & W met = 0 if W NOT met
Probability of a criticality at state i
Probability of reaching normal state if NO additional criticality occurs at state i
Probability of reaching normal state if ANY additional criticality occurs at state i
Probability of reaching the normal state from state i