SlideShare a Scribd company logo
1 of 20
Collaborative Recommender
Systems for Building
Automation
Michael LeMay, Jason J. Haas, and
Carl A. Gunter
University of Illinois
• Motivation: Future Building Automation Systems
(BASs) will support a wide variety of control
algorithms
– Managers may not be able to determine which
algorithm is the best on their own
• Approach: Use recommender system to help
managers share opinions and quantitative
comparisons of algorithms, to result in optimal
performance
Overview
• Siebel Center for Computer
Science
–Centralized system permits
monitoring and control of:
• HVAC
• Card-swipe door locks
• Motion sensors
• Lighting
Sample Industrial BAS
• Analyze electrical consumption at
a few key points (e.g. each circuit
breaker) to determine the states
of the appliances attached to
those points
• Many possible algorithms…
– Threshold-based (incrementally
adjust appliance states based on
energy consumption changes)
– 0-1 knapsack (computationally
expensive)
Non-Intrusive Load Monitoring
• Increased occupant comfort relative to
configuration effort
• Decreased energy consumption
• Decreased energy cost for a given level of
consumption
• Better visibility into electrical consumption
Possible BAS Benefits
• BASs could deployed in a variety of
environments:
– Private homes
– Hotels
– Retail stores
– Warehouses
– Office buildings
BAS Applicability
• Private home with working parents and kids in
school:
– Occupied mostly from evenings through mornings and on
weekends
– Occasional guests with special requirements (e.g. extra
heat or cold, use of guest room)
• Private home with homemaker and kids at home:
– Occupied most of the day and night
• Hotel
– Similar to first scenario, but occupants change every day
or so and housekeepers stop by in middle of day
Environmental Characteristics
• Retail Store
– Uniformly occupied for large portions of day by large
quantities of people
– Certain parts of store have special requirements (e.g.
freezer section should be colder than other aisles)
• Warehouses
– Sparsely occupied throughout the business day by
highly-active people specially-equipped to operate in
environment (e.g. wearing coats)
– Particular sections may have special requirements,
such as a small side-office
Environmental Characteristics
(cont.)
• Office buildings
– Segmented into many small spaces with varying
requirements that are occupied throughout the
business day by an infrequently-changing set of
people.
– A few spaces such as conference rooms will be
unoccupied for many parts of the day, and have
various groups of people in them in other parts of the
day
Environmental Characteristics
(cont.)
• Lighting algorithm that turns off lights when
motion has not been detected for certain period
of time:
– In office: May turn off lights when person is relatively
still, causing annoyance.
– In retail store: Highly-effective, since shoppers rarely
stop moving
• NILM algorithm that operates using thresholds:
– Will be more effective in an environment with
appliances that can be turned on and off than one
with variable-speed motors, for example.
Effect on Control Algorithm
Effectiveness
• Example #1:
– Motion sensor detects occupant getting up in morning
– BAS turns on hallway and kitchen lights
– Not effective in a hotel where different occupants have different habits
• Example #2:
– Motion sensor detects occupant in room, and subsequently turns on the
lights to their maximum intensity.
– The next day, when an occupant re-enters the room, the BAS
automatically turns the lights to 2/3 of their maximum intensity.
– The occupant immediately increases the intensity to the maximum.
– The next day, the BAS uses 5/6 of maximum intensity, and the occupant is
content, as indicated by the fact that they do not subsequently increase
the intensity.
– Again, not effective in environment with rapidly-changing sets of
occupants with different preferences
More Examples
• Content-dependent: Recommendations made based on
similarity of new items to items previously rated by user
• Content-independent: Recommender unaware of
characteristics of items being recommended, except their
ratings from other users
– E.g. Social filtering: Generate new rating based on rating of
others, giving more weight to ratings from “similar” users
• Amazon probably uses a hybrid: Recommends items
similar to items I purchased previously, plus items
purchased by other people with similar purchase histories.
Recommender Systems
• Evaluate similarity
of building
managers:
• Generate
prediction:
Social Filtering
• Use a recommender system to recommend BAS
algorithms to building managers
• Challenging to determine in general how similar
the “contents” of algorithms are, so social
filtering is a better choice in the context of BAS
algorithms
– Building managers fill out a survey characterizing
their buildings so that their recommendations are
weighted more highly with managers of similar
buildings.
Approach
CollaborVation Architecture
Collaborative
Recommender
Animated Operational Overview
Energy
Modeler
Appliance
Usage
Detector
Occupancy
Detector Setpoint
Generator
Discomfort
Predictor
Energy
Usage
Predictor
Energy
Cost
Predictor
X10 USB
Transceiver
• We used the Duine recommender software for Java
to rate individual module implementations
• Provides implementations for several recommender
algorithms: User Average, TopN Deviation, etc.
• We selected Social Filtering
– All ratings of a particular algorithm are weighted by the
similarity between the building considering the algorithm
and the building that generated the rating.
– The weighted average of the ratings is the predicted
rating of the algorithm in the “querying” building.
Recommender System Prototype
• Five buildings: Two apartments, two small retail
stores, one industrial plant with a small office.
• Renters in apartments rate NILM algorithm #1
highly, and NILM algorithm #2 poorly
• Owner of retail store #1 rates NILM algorithm #2
poorly, and NILM algorithm #1 highly
• Owner of industrial plant rates both equally.
• Manager of store #2 requests a rating. The result?
• NILM algorithm #2 ranked lower than NILM
algorithm #1, but the rating is slightly higher than
the one provided by store #1
Recommender Example Scenario
• BAS algorithms may become sufficiently numerous
and complex that managers have difficulty
independently selecting the best ones for their
applications
• Recommender systems may help managers to select
appropriate algorithms
• A loosely-coupled blackboard architecture permits
BAS algorithms to be dynamically swapped when
changes are recommended
• All technologies necessary for implementation are
readily-available and reliable
Conclusion
• http://seclab.uiuc.edu
Thank You!

More Related Content

Viewers also liked

Soa amsterdam-restws-pautasso-talk
Soa amsterdam-restws-pautasso-talkSoa amsterdam-restws-pautasso-talk
Soa amsterdam-restws-pautasso-talkAravindharamanan S
 
Yoda an accurate and scalable web based recommendation systems
Yoda an accurate and scalable web based recommendation systemsYoda an accurate and scalable web based recommendation systems
Yoda an accurate and scalable web based recommendation systemsAravindharamanan S
 
Ppt bigdataanalyticsfinanceprofessionals-big data, business analytics and fin...
Ppt bigdataanalyticsfinanceprofessionals-big data, business analytics and fin...Ppt bigdataanalyticsfinanceprofessionals-big data, business analytics and fin...
Ppt bigdataanalyticsfinanceprofessionals-big data, business analytics and fin...Aravindharamanan S
 
From home sql_server_tutorials
From home sql_server_tutorialsFrom home sql_server_tutorials
From home sql_server_tutorialsAravindharamanan S
 
Toward the next generation of recommender systems
Toward the next generation of recommender systemsToward the next generation of recommender systems
Toward the next generation of recommender systemsAravindharamanan S
 
Visual studio-2012-product-guide
Visual studio-2012-product-guideVisual studio-2012-product-guide
Visual studio-2012-product-guideAravindharamanan S
 

Viewers also liked (11)

Soa amsterdam-restws-pautasso-talk
Soa amsterdam-restws-pautasso-talkSoa amsterdam-restws-pautasso-talk
Soa amsterdam-restws-pautasso-talk
 
Yoda an accurate and scalable web based recommendation systems
Yoda an accurate and scalable web based recommendation systemsYoda an accurate and scalable web based recommendation systems
Yoda an accurate and scalable web based recommendation systems
 
Ppt bigdataanalyticsfinanceprofessionals-big data, business analytics and fin...
Ppt bigdataanalyticsfinanceprofessionals-big data, business analytics and fin...Ppt bigdataanalyticsfinanceprofessionals-big data, business analytics and fin...
Ppt bigdataanalyticsfinanceprofessionals-big data, business analytics and fin...
 
From home sql_server_tutorials
From home sql_server_tutorialsFrom home sql_server_tutorials
From home sql_server_tutorials
 
Toward the next generation of recommender systems
Toward the next generation of recommender systemsToward the next generation of recommender systems
Toward the next generation of recommender systems
 
Recommender systems
Recommender systemsRecommender systems
Recommender systems
 
Xml+messaging+with+soap
Xml+messaging+with+soapXml+messaging+with+soap
Xml+messaging+with+soap
 
Soap binding survey
Soap binding surveySoap binding survey
Soap binding survey
 
Visual studio-2012-product-guide
Visual studio-2012-product-guideVisual studio-2012-product-guide
Visual studio-2012-product-guide
 
Android dev 3
Android dev 3Android dev 3
Android dev 3
 
Android programming-basics
Android programming-basicsAndroid programming-basics
Android programming-basics
 

Similar to Collaborative recommender systems for building automation

4. Luca Ferrarini (POLIMI, Italy) - Pilot Case 1: The Reality of Working with...
4. Luca Ferrarini (POLIMI, Italy) - Pilot Case 1: The Reality of Working with...4. Luca Ferrarini (POLIMI, Italy) - Pilot Case 1: The Reality of Working with...
4. Luca Ferrarini (POLIMI, Italy) - Pilot Case 1: The Reality of Working with...Cassandra Project
 
esci-61-pv-system-design-and-sizing-slides.pdf
esci-61-pv-system-design-and-sizing-slides.pdfesci-61-pv-system-design-and-sizing-slides.pdf
esci-61-pv-system-design-and-sizing-slides.pdfRiazAhmad205274
 
Catalogo dialux ro beu-part1-2
Catalogo dialux ro beu-part1-2Catalogo dialux ro beu-part1-2
Catalogo dialux ro beu-part1-2Jose Suarez
 
Whole Building Modeling Using OpenStudio
Whole Building Modeling Using OpenStudioWhole Building Modeling Using OpenStudio
Whole Building Modeling Using OpenStudioenergytrustor
 
Optimizing Your AWS Applications and Usage to Reduce Costs
Optimizing Your AWS Applications and Usage to Reduce CostsOptimizing Your AWS Applications and Usage to Reduce Costs
Optimizing Your AWS Applications and Usage to Reduce CostsAmazon Web Services
 
Occupant Engaged - A ZEB control concept for engaging occupants
Occupant Engaged -  A ZEB control concept for engaging occupantsOccupant Engaged -  A ZEB control concept for engaging occupants
Occupant Engaged - A ZEB control concept for engaging occupantsShanti Pless
 
SYBSC IT SEM IV EMBEDDED SYSTEMS UNIT I Characteristics and Quality Attribute...
SYBSC IT SEM IV EMBEDDED SYSTEMS UNIT I Characteristics and Quality Attribute...SYBSC IT SEM IV EMBEDDED SYSTEMS UNIT I Characteristics and Quality Attribute...
SYBSC IT SEM IV EMBEDDED SYSTEMS UNIT I Characteristics and Quality Attribute...Arti Parab Academics
 
Optimizing your cloud
Optimizing your cloudOptimizing your cloud
Optimizing your cloud2nd Watch
 
AWS Sydney Summit 2013 - Optimizing AWS Applications and Usage to Reduce Costs
AWS Sydney Summit 2013 - Optimizing AWS Applications and Usage to Reduce CostsAWS Sydney Summit 2013 - Optimizing AWS Applications and Usage to Reduce Costs
AWS Sydney Summit 2013 - Optimizing AWS Applications and Usage to Reduce CostsAmazon Web Services
 
CIS375 Interaction Designs Chapter15
CIS375 Interaction Designs Chapter15CIS375 Interaction Designs Chapter15
CIS375 Interaction Designs Chapter15Dr. Ahmed Al Zaidy
 
Design principles & quality factors
Design principles & quality factorsDesign principles & quality factors
Design principles & quality factorsAalia Barbe
 
Using Metrics for Fun, Developing with the KV Store + Javascript & News from ...
Using Metrics for Fun, Developing with the KV Store + Javascript & News from ...Using Metrics for Fun, Developing with the KV Store + Javascript & News from ...
Using Metrics for Fun, Developing with the KV Store + Javascript & News from ...Harry McLaren
 
05. performance-concepts-26-slides
05. performance-concepts-26-slides05. performance-concepts-26-slides
05. performance-concepts-26-slidesMuhammad Ahad
 
activity_based_costing power point presentation
activity_based_costing power point presentationactivity_based_costing power point presentation
activity_based_costing power point presentationChamodiBandara1
 
AWS Summit Berlin 2013 - Optimizing your AWS applications and usage to reduce...
AWS Summit Berlin 2013 - Optimizing your AWS applications and usage to reduce...AWS Summit Berlin 2013 - Optimizing your AWS applications and usage to reduce...
AWS Summit Berlin 2013 - Optimizing your AWS applications and usage to reduce...AWS Germany
 
Ncerc rlmca202 adm m4 ssm
Ncerc rlmca202 adm m4 ssmNcerc rlmca202 adm m4 ssm
Ncerc rlmca202 adm m4 ssmssmarar
 
6. Anastasios Bakaoukas (COVUNI, UK) - Pilot Case 3: A Lighting Product Bench...
6. Anastasios Bakaoukas (COVUNI, UK) - Pilot Case 3: A Lighting Product Bench...6. Anastasios Bakaoukas (COVUNI, UK) - Pilot Case 3: A Lighting Product Bench...
6. Anastasios Bakaoukas (COVUNI, UK) - Pilot Case 3: A Lighting Product Bench...Cassandra Project
 
Is accurate system-level power measurement challenging? Check this out!
Is accurate system-level power measurement challenging? Check this out!Is accurate system-level power measurement challenging? Check this out!
Is accurate system-level power measurement challenging? Check this out!Deepak Shankar
 

Similar to Collaborative recommender systems for building automation (20)

4. Luca Ferrarini (POLIMI, Italy) - Pilot Case 1: The Reality of Working with...
4. Luca Ferrarini (POLIMI, Italy) - Pilot Case 1: The Reality of Working with...4. Luca Ferrarini (POLIMI, Italy) - Pilot Case 1: The Reality of Working with...
4. Luca Ferrarini (POLIMI, Italy) - Pilot Case 1: The Reality of Working with...
 
esci-61-pv-system-design-and-sizing-slides.pdf
esci-61-pv-system-design-and-sizing-slides.pdfesci-61-pv-system-design-and-sizing-slides.pdf
esci-61-pv-system-design-and-sizing-slides.pdf
 
Catalogo dialux ro beu-part1-2
Catalogo dialux ro beu-part1-2Catalogo dialux ro beu-part1-2
Catalogo dialux ro beu-part1-2
 
Whole Building Modeling Using OpenStudio
Whole Building Modeling Using OpenStudioWhole Building Modeling Using OpenStudio
Whole Building Modeling Using OpenStudio
 
Optimizing Your AWS Applications and Usage to Reduce Costs
Optimizing Your AWS Applications and Usage to Reduce CostsOptimizing Your AWS Applications and Usage to Reduce Costs
Optimizing Your AWS Applications and Usage to Reduce Costs
 
Occupant Engaged - A ZEB control concept for engaging occupants
Occupant Engaged -  A ZEB control concept for engaging occupantsOccupant Engaged -  A ZEB control concept for engaging occupants
Occupant Engaged - A ZEB control concept for engaging occupants
 
SYBSC IT SEM IV EMBEDDED SYSTEMS UNIT I Characteristics and Quality Attribute...
SYBSC IT SEM IV EMBEDDED SYSTEMS UNIT I Characteristics and Quality Attribute...SYBSC IT SEM IV EMBEDDED SYSTEMS UNIT I Characteristics and Quality Attribute...
SYBSC IT SEM IV EMBEDDED SYSTEMS UNIT I Characteristics and Quality Attribute...
 
Optimizing your cloud
Optimizing your cloudOptimizing your cloud
Optimizing your cloud
 
AWS Sydney Summit 2013 - Optimizing AWS Applications and Usage to Reduce Costs
AWS Sydney Summit 2013 - Optimizing AWS Applications and Usage to Reduce CostsAWS Sydney Summit 2013 - Optimizing AWS Applications and Usage to Reduce Costs
AWS Sydney Summit 2013 - Optimizing AWS Applications and Usage to Reduce Costs
 
CIS375 Interaction Designs Chapter15
CIS375 Interaction Designs Chapter15CIS375 Interaction Designs Chapter15
CIS375 Interaction Designs Chapter15
 
Design principles & quality factors
Design principles & quality factorsDesign principles & quality factors
Design principles & quality factors
 
Using Metrics for Fun, Developing with the KV Store + Javascript & News from ...
Using Metrics for Fun, Developing with the KV Store + Javascript & News from ...Using Metrics for Fun, Developing with the KV Store + Javascript & News from ...
Using Metrics for Fun, Developing with the KV Store + Javascript & News from ...
 
05. performance-concepts-26-slides
05. performance-concepts-26-slides05. performance-concepts-26-slides
05. performance-concepts-26-slides
 
activity_based_costing power point presentation
activity_based_costing power point presentationactivity_based_costing power point presentation
activity_based_costing power point presentation
 
Prototype Model
Prototype ModelPrototype Model
Prototype Model
 
AWS Summit Berlin 2013 - Optimizing your AWS applications and usage to reduce...
AWS Summit Berlin 2013 - Optimizing your AWS applications and usage to reduce...AWS Summit Berlin 2013 - Optimizing your AWS applications and usage to reduce...
AWS Summit Berlin 2013 - Optimizing your AWS applications and usage to reduce...
 
Ncerc rlmca202 adm m4 ssm
Ncerc rlmca202 adm m4 ssmNcerc rlmca202 adm m4 ssm
Ncerc rlmca202 adm m4 ssm
 
6. Anastasios Bakaoukas (COVUNI, UK) - Pilot Case 3: A Lighting Product Bench...
6. Anastasios Bakaoukas (COVUNI, UK) - Pilot Case 3: A Lighting Product Bench...6. Anastasios Bakaoukas (COVUNI, UK) - Pilot Case 3: A Lighting Product Bench...
6. Anastasios Bakaoukas (COVUNI, UK) - Pilot Case 3: A Lighting Product Bench...
 
Is accurate system-level power measurement challenging? Check this out!
Is accurate system-level power measurement challenging? Check this out!Is accurate system-level power measurement challenging? Check this out!
Is accurate system-level power measurement challenging? Check this out!
 
Energy management system
Energy management systemEnergy management system
Energy management system
 

Recently uploaded

Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...srsj9000
 
Internship report on mechanical engineering
Internship report on mechanical engineeringInternship report on mechanical engineering
Internship report on mechanical engineeringmalavadedarshan25
 
Microscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptxMicroscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptxpurnimasatapathy1234
 
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝soniya singh
 
College Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service NashikCollege Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service NashikCall Girls in Nagpur High Profile
 
Past, Present and Future of Generative AI
Past, Present and Future of Generative AIPast, Present and Future of Generative AI
Past, Present and Future of Generative AIabhishek36461
 
SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )Tsuyoshi Horigome
 
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...VICTOR MAESTRE RAMIREZ
 
microprocessor 8085 and its interfacing
microprocessor 8085  and its interfacingmicroprocessor 8085  and its interfacing
microprocessor 8085 and its interfacingjaychoudhary37
 
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdfCCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdfAsst.prof M.Gokilavani
 
Heart Disease Prediction using machine learning.pptx
Heart Disease Prediction using machine learning.pptxHeart Disease Prediction using machine learning.pptx
Heart Disease Prediction using machine learning.pptxPoojaBan
 
Software and Systems Engineering Standards: Verification and Validation of Sy...
Software and Systems Engineering Standards: Verification and Validation of Sy...Software and Systems Engineering Standards: Verification and Validation of Sy...
Software and Systems Engineering Standards: Verification and Validation of Sy...VICTOR MAESTRE RAMIREZ
 
HARMONY IN THE HUMAN BEING - Unit-II UHV-2
HARMONY IN THE HUMAN BEING - Unit-II UHV-2HARMONY IN THE HUMAN BEING - Unit-II UHV-2
HARMONY IN THE HUMAN BEING - Unit-II UHV-2RajaP95
 
IVE Industry Focused Event - Defence Sector 2024
IVE Industry Focused Event - Defence Sector 2024IVE Industry Focused Event - Defence Sector 2024
IVE Industry Focused Event - Defence Sector 2024Mark Billinghurst
 
High Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur EscortsHigh Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur EscortsCall Girls in Nagpur High Profile
 
Biology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptxBiology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptxDeepakSakkari2
 
chaitra-1.pptx fake news detection using machine learning
chaitra-1.pptx  fake news detection using machine learningchaitra-1.pptx  fake news detection using machine learning
chaitra-1.pptx fake news detection using machine learningmisbanausheenparvam
 

Recently uploaded (20)

Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
 
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptxExploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
 
Internship report on mechanical engineering
Internship report on mechanical engineeringInternship report on mechanical engineering
Internship report on mechanical engineering
 
Microscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptxMicroscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptx
 
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
 
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝
 
College Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service NashikCollege Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
 
Past, Present and Future of Generative AI
Past, Present and Future of Generative AIPast, Present and Future of Generative AI
Past, Present and Future of Generative AI
 
SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )
 
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
 
microprocessor 8085 and its interfacing
microprocessor 8085  and its interfacingmicroprocessor 8085  and its interfacing
microprocessor 8085 and its interfacing
 
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdfCCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
 
Heart Disease Prediction using machine learning.pptx
Heart Disease Prediction using machine learning.pptxHeart Disease Prediction using machine learning.pptx
Heart Disease Prediction using machine learning.pptx
 
Software and Systems Engineering Standards: Verification and Validation of Sy...
Software and Systems Engineering Standards: Verification and Validation of Sy...Software and Systems Engineering Standards: Verification and Validation of Sy...
Software and Systems Engineering Standards: Verification and Validation of Sy...
 
HARMONY IN THE HUMAN BEING - Unit-II UHV-2
HARMONY IN THE HUMAN BEING - Unit-II UHV-2HARMONY IN THE HUMAN BEING - Unit-II UHV-2
HARMONY IN THE HUMAN BEING - Unit-II UHV-2
 
IVE Industry Focused Event - Defence Sector 2024
IVE Industry Focused Event - Defence Sector 2024IVE Industry Focused Event - Defence Sector 2024
IVE Industry Focused Event - Defence Sector 2024
 
High Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur EscortsHigh Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Meera Call 7001035870 Meet With Nagpur Escorts
 
young call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Service
young call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Serviceyoung call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Service
young call girls in Rajiv Chowk🔝 9953056974 🔝 Delhi escort Service
 
Biology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptxBiology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptx
 
chaitra-1.pptx fake news detection using machine learning
chaitra-1.pptx  fake news detection using machine learningchaitra-1.pptx  fake news detection using machine learning
chaitra-1.pptx fake news detection using machine learning
 

Collaborative recommender systems for building automation

  • 1. Collaborative Recommender Systems for Building Automation Michael LeMay, Jason J. Haas, and Carl A. Gunter University of Illinois
  • 2. • Motivation: Future Building Automation Systems (BASs) will support a wide variety of control algorithms – Managers may not be able to determine which algorithm is the best on their own • Approach: Use recommender system to help managers share opinions and quantitative comparisons of algorithms, to result in optimal performance Overview
  • 3. • Siebel Center for Computer Science –Centralized system permits monitoring and control of: • HVAC • Card-swipe door locks • Motion sensors • Lighting Sample Industrial BAS
  • 4. • Analyze electrical consumption at a few key points (e.g. each circuit breaker) to determine the states of the appliances attached to those points • Many possible algorithms… – Threshold-based (incrementally adjust appliance states based on energy consumption changes) – 0-1 knapsack (computationally expensive) Non-Intrusive Load Monitoring
  • 5. • Increased occupant comfort relative to configuration effort • Decreased energy consumption • Decreased energy cost for a given level of consumption • Better visibility into electrical consumption Possible BAS Benefits
  • 6. • BASs could deployed in a variety of environments: – Private homes – Hotels – Retail stores – Warehouses – Office buildings BAS Applicability
  • 7. • Private home with working parents and kids in school: – Occupied mostly from evenings through mornings and on weekends – Occasional guests with special requirements (e.g. extra heat or cold, use of guest room) • Private home with homemaker and kids at home: – Occupied most of the day and night • Hotel – Similar to first scenario, but occupants change every day or so and housekeepers stop by in middle of day Environmental Characteristics
  • 8. • Retail Store – Uniformly occupied for large portions of day by large quantities of people – Certain parts of store have special requirements (e.g. freezer section should be colder than other aisles) • Warehouses – Sparsely occupied throughout the business day by highly-active people specially-equipped to operate in environment (e.g. wearing coats) – Particular sections may have special requirements, such as a small side-office Environmental Characteristics (cont.)
  • 9. • Office buildings – Segmented into many small spaces with varying requirements that are occupied throughout the business day by an infrequently-changing set of people. – A few spaces such as conference rooms will be unoccupied for many parts of the day, and have various groups of people in them in other parts of the day Environmental Characteristics (cont.)
  • 10. • Lighting algorithm that turns off lights when motion has not been detected for certain period of time: – In office: May turn off lights when person is relatively still, causing annoyance. – In retail store: Highly-effective, since shoppers rarely stop moving • NILM algorithm that operates using thresholds: – Will be more effective in an environment with appliances that can be turned on and off than one with variable-speed motors, for example. Effect on Control Algorithm Effectiveness
  • 11. • Example #1: – Motion sensor detects occupant getting up in morning – BAS turns on hallway and kitchen lights – Not effective in a hotel where different occupants have different habits • Example #2: – Motion sensor detects occupant in room, and subsequently turns on the lights to their maximum intensity. – The next day, when an occupant re-enters the room, the BAS automatically turns the lights to 2/3 of their maximum intensity. – The occupant immediately increases the intensity to the maximum. – The next day, the BAS uses 5/6 of maximum intensity, and the occupant is content, as indicated by the fact that they do not subsequently increase the intensity. – Again, not effective in environment with rapidly-changing sets of occupants with different preferences More Examples
  • 12. • Content-dependent: Recommendations made based on similarity of new items to items previously rated by user • Content-independent: Recommender unaware of characteristics of items being recommended, except their ratings from other users – E.g. Social filtering: Generate new rating based on rating of others, giving more weight to ratings from “similar” users • Amazon probably uses a hybrid: Recommends items similar to items I purchased previously, plus items purchased by other people with similar purchase histories. Recommender Systems
  • 13. • Evaluate similarity of building managers: • Generate prediction: Social Filtering
  • 14. • Use a recommender system to recommend BAS algorithms to building managers • Challenging to determine in general how similar the “contents” of algorithms are, so social filtering is a better choice in the context of BAS algorithms – Building managers fill out a survey characterizing their buildings so that their recommendations are weighted more highly with managers of similar buildings. Approach
  • 16. Animated Operational Overview Energy Modeler Appliance Usage Detector Occupancy Detector Setpoint Generator Discomfort Predictor Energy Usage Predictor Energy Cost Predictor X10 USB Transceiver
  • 17. • We used the Duine recommender software for Java to rate individual module implementations • Provides implementations for several recommender algorithms: User Average, TopN Deviation, etc. • We selected Social Filtering – All ratings of a particular algorithm are weighted by the similarity between the building considering the algorithm and the building that generated the rating. – The weighted average of the ratings is the predicted rating of the algorithm in the “querying” building. Recommender System Prototype
  • 18. • Five buildings: Two apartments, two small retail stores, one industrial plant with a small office. • Renters in apartments rate NILM algorithm #1 highly, and NILM algorithm #2 poorly • Owner of retail store #1 rates NILM algorithm #2 poorly, and NILM algorithm #1 highly • Owner of industrial plant rates both equally. • Manager of store #2 requests a rating. The result? • NILM algorithm #2 ranked lower than NILM algorithm #1, but the rating is slightly higher than the one provided by store #1 Recommender Example Scenario
  • 19. • BAS algorithms may become sufficiently numerous and complex that managers have difficulty independently selecting the best ones for their applications • Recommender systems may help managers to select appropriate algorithms • A loosely-coupled blackboard architecture permits BAS algorithms to be dynamically swapped when changes are recommended • All technologies necessary for implementation are readily-available and reliable Conclusion

Editor's Notes

  1. Benefits of DR: For consumers, reduced costs; For the ESPs, reduced provisioning requirements For regulators, improved reliability.
  2. Advanced: The DoE report does not imagine that residential consumers know enough about their usage to take advantage of DR. Indeed they are skeptical of its benefits for medium sized customers and recommend it only for large customers. Limited: They don’t use advanced meters (yet) and there is no support for automation. A CA study showed that participants reduced load 13 percent on average, but as much as 27 percent when price signals were coupled with automated controls such as controllable thermostats. (Charles River Associates. Impact Evaluation of the California Statewide Pricing Pilot: Final Report. March 16, 2005.) Remarks on an Illinois program in Alexander07. Is this program related to Power Smart Pricing? One of the more intriguing real time pricing pilot programs targeted to residential customers has been operated on a pilot basis for three years in Chicago, Illinois by Community Energy Cooperative, a local community organization in cooperation with Commonwealth Edison (ComEd). Participating customers were provided a new interval meter and charged hourly prices for electricity based on the PJM ComEd zone day ahead hourly locational marginal price. These meters were not “smart meters” because they lacked two-way communications between the meter and the utility. Rather, the meter recorded usage on hourly intervals, but was manually read by an electronic probe and the usage information downloaded to the utility billing system. Customers were informed of these day ahead hourly prices via an email communication or they accessed the same information on the Community Energy Cooperative website. When an hourly price was slated to hit a predetermined trigger, customers were specially notified of that event.
  3. Talk about MGA here…
  4. Break out final point