Orientation & Selected Guest Lecture Models Nathaniel Osgood - - PowerPoint PPT Presentation

Orientation & Selected Guest Lecture Models

Nathaniel Osgood Agent-Based Modeling Bootcamp for Health Researchers August 22, 2011

Goals of Bootcamp

• To expose participants to basics of building agent-

based models (ABMs) in AnyLogic

• To convey a sense as to the capabilities, strengths and

weaknesses of agent-based modeling

• To train AnyLogic users so that they are comfortable

running, understanding many elements of, and performing small modifications on ABMs

• To provide aspiring modelers with hands-on familiarity

with the capabilities & functions of AnyLogic and sufficient exposure to Java programming to navigate within models

• To help create informed consumers of Agent-based

models

Non-Goals of the Bootcamp

• To serve as a comprehensive survey of Agent-Based

modeling

• To give a thorough conceptual or theoretic framework

for understanding and analyzing Agent-Based Models

• To review the tradeoffs between AnyLogic & other ABM

platforms

• To create a set of fully proficient Agent-Based modelers
• To render participants into Java programmers
• To sell participants on the superiority of AnyLogic, or

the desirability of buying AnyLogic licenses

Ways Seeking to Resemble a Bootcamp….

• Intense
• Long hours
• Willingness to push limits
• No frills
• Development of sense of camaraderie
• Tolerance of a loud & (technologically)

aggressive instructor

• Thorough exercise of (brain) muscles
• Looking back at end

– Sense of accomplishment – New appreciation of what are capable of

Ways Seeking to Avoid Resembling a Bootcamp….

• No derogatory language or demeaning humour
• Breaking down barrier between instructor/TAs &

participants

• Willingness of participants to speak up with

– Questions – Suggestions – Objections

• Rich dialogue within room
• Inclusive atmosphere

Notes on the Schedule

• Each day, schedule incorporates

– Many (primarily) hands-on plenary lectures – Guest lectures highlighting ABM applications in health – Parallel sessions w/more advanced material/exercises – Breaks – Discussion periods

• Color coded by level of material & nature of session
• The schedule is not immutable: Subject to change

based on

– Sense of whethre more breaks are required – Student interest & feedback – Perception of comfort with & absorption of material

Teaching Assistants

• Yuan Tian (Head TA, Computer Science)

– Formal Health & CS training; Tuberculosis prevention & control application, contact tracing protocol simulation, TB&Smoking

• Amy Gao (Computer Science)

– Diabetes Epidemiology, Diabetic & ESRD, Health Care Cost estimation with models, leveraging database tie-ins

• Yudi Xue (Computer Science)

– AnyLogic “under the hood”, agent subtyping, software for modeling process facilitation via versioning&collaboration

• Anqi Dong (Walter Murray Collegiate Institute)

– ABM for evaluating effectiveness of treatment prioritization strategies, evaluating Bayesian inferencing

• Weicheng Qian (Computer Science)

– Agent mobility modeling, Bayesian estimation & filtering

Other Participants w/ABM Exposure

• David Vickers

– Immuno-epidemiological modeling with ABM- System Dynamics Hybrids, STIs, N. meningitidis

• Mohammad Hashemian

– Incorporation of person-place & person-person sensor micro-contact data into ABM simulations

• Cheryl Waldner

– Help advise on construction of CWD model, Zoonoses, Veterinary Medicine, H1N1 modeling

Teaching Assistant Roles

• Assisting participants with ad-hoc questions &

modeling difficulties (e.g. with exercises)

• Discussing Details of presented & example

models

• Brainstorming on ABM applications
• Advising on AnyLogic capabilities, functions &

limitations

• Consultations on modeling topics
• Guest lectures

Practical Essentials

• Java & AnyLogic Sessions held in Spinks S386
• Longer consultations with TAs/instructor can be

held in Spinks S371

• Washrooms & water fountain are down the hall

to left

• Quiet lounge area (e.g. for phone conversations)

is to the right

• USB Stick contains

– Binder contents – Sample models – Preliminary Presentations

• Red Dots Identify TAs, Yellow Dots all Volunteers

Food

• “Bootcamp Breakfasts” and snacks will be

provided in hallway outside S320

• Lunches (& lunch talks) in Faculty Club

– Some downstairs, some upstairs

• Dinners on your own (restaurants suggestions

in binder & on USB drive)

IT Essentials

• Your user name & password should be in the back of

your nametag

– Please log in with this information throughout the workshop

• Once logged in to a machine, go to windows

explorer & “X:” drive to find files from instructor

– Preliminary Lecture Slides/Exercises/Example Models (take home & not take home)/Binder content

• Your files will be stored with your account, not with

the particular machine

– But try to avoid relying on files stored to desktop (store to “My documents” instead).

• WiFi Wireless networks are available

– Use “guest” as network (SSID), “visitor” as both user name & password

Finding S371 & S386 (Ignore Green Dots!) S320

I am here! S371 (for Consultations with TAs & Instructor) Path from S320 to S371

S371 S386

S386 (for AnyLogic & Java Tutorials)

A First Encounter With Anylogic: Modifying A Simple Sample Model Building a Minimalist Network-Based Model Framework A Simple Network Based Infection Spread Model Building A Minimalist Two-Population Model Framework Incorporating Attribute Heterogeneity Among Agents A Simple Debugging Exercise

Exercises: Prerequisites

Legend

Recommended Required Introductory Basic Intermediate

Target Skill Level Strength of Prerequisite Relationships

Glimpse of Guest Lectures

(Primarily over Lunch)

• TB
• CWD
• ESRD
• Treatment Prioritization
• ABMs for Reliability of Inferencing
• HPV & Smoking
• Alzheimer’s Disease
• SILVER

Tuberculosis Spread, Prevention & Control

(Earlier Version)

16

Chronic Wasting Disease

Health & Cost Implications of Diabetic ESRD

HPV & Smoking

SILVER Versioning & Collaboration System

Free virus κy uv β1xv Uninfected Cells Infected Cells β2xy Activated CD8+ cells Memory CTLs Clonally- expanding CD8+ cells ay+dx+pφzy dx ay cyw cθw pφzy hz z0/m bw z8/m

Within Host Model

Network Embedded Individuals

Uninfected Cells Infected Cells Virus Load Uninfected Cell Replentishment New Cell Infections Uninfected Cell death Infected Cell Death Virion Production From Infected Cells Virion Clearance Uninfected Cell Replentishment Rate Mean Infected Cell Lifetime Mean Uninfected Cell Lifetime Mean Virion Lifetime Likelihood Density of Infection by Single Virion Per Infected CellVirion Production Rate Virion Production Rate Per Contact Virions Rate 1 Person Mean Viral Load <Population Size> Mean Uninfected Cells Mean Infected Cells <Population Size> <Population Size> Mean of Viral Load

• f Neighbors

CTLs immune response to infected cells CTL turnover CTL responsiveness Mean CTL lifespan infected cell death by CTLs rate which infected cells are killed by CTLs Virion Production Rate if Non Quantized Infection