Presentation
Integrative Modeling and Simulation of Human Behavior and Human-Machine Systems With the Queuing Network Architecture
DescriptionAs in any mature field of science and engineering, human factors engineering (HFE) needs integrative computational frameworks that unify a large and diverse range of theories and methods. For more than two decades, the workshop instructor and his collaborators have been developing the Queuing Network (QN) Architecture, using it to unify many existing theories and methods, and applying it to a variety of HFE applications. This workshop describes and illustrates how to use the QN Architecture to evaluate a wide range of theories and methods, and to model and simulate a wide range of human behavior and human-machine systems at various levels of granularity.
Step-by-step and alternating between lectures and software demos, this workshop describes the family of QN models that account for phenomena from the micro-neural level to meso-individual level to macro-team or multi-agent level. These "model siblings" all speak the same QN language and share the same QN ideas, but each has its specialties and companions/partners (e.g., AI, ML, Optimization): (1) “QN-Reaction Time” includes many RT models as special cases; (2) “QN-RT-Accuracy” bridges QN with Diffusion/Accumulator models to account for Speed-Accuracy Tradeoff (SAT), etc.; (3) “QN-MHP” integrates QN with Model-Human-Processor (MHP) to model procedure-based multi-task performance; (4) “QN-ACT-R” integrates QN with ACT-R by implementing and visualizing ACT-R’s modules and buffers as QN servers to model multitask performance involving complex cognition; (5) “QN-MBS (Mind-Body-System)” extends the QN architecture to include body parts; (6) “QN-Neural” models neurophysiological phenomena; (7) “QN-MPMM” (Multi-Person Multi-Machine) treats MPMM systems as hierarchically-organized larger QNs. More siblings are in various stages of incubation (e.g., QN-Control, QN-NSEEV and QN-ACES).
QN models were originally coded in commercial simulation software but are now being implemented in Python with easy-to-use interfaces. Model “Users” do not need to know Python. “Developers” can delve into and modify the codes.
Students/researchers/practitioners of any level of expertise can benefit from this workshop: learn how to use the QN methods/software (“get a feel for it”), incorporate it in dissertation/theoretical/applied research or product evaluation, and/or extend/challenge the models. A laptop is optional for hearing the lectures and seeing the demos but needed for accessing electronic documents.
Step-by-step and alternating between lectures and software demos, this workshop describes the family of QN models that account for phenomena from the micro-neural level to meso-individual level to macro-team or multi-agent level. These "model siblings" all speak the same QN language and share the same QN ideas, but each has its specialties and companions/partners (e.g., AI, ML, Optimization): (1) “QN-Reaction Time” includes many RT models as special cases; (2) “QN-RT-Accuracy” bridges QN with Diffusion/Accumulator models to account for Speed-Accuracy Tradeoff (SAT), etc.; (3) “QN-MHP” integrates QN with Model-Human-Processor (MHP) to model procedure-based multi-task performance; (4) “QN-ACT-R” integrates QN with ACT-R by implementing and visualizing ACT-R’s modules and buffers as QN servers to model multitask performance involving complex cognition; (5) “QN-MBS (Mind-Body-System)” extends the QN architecture to include body parts; (6) “QN-Neural” models neurophysiological phenomena; (7) “QN-MPMM” (Multi-Person Multi-Machine) treats MPMM systems as hierarchically-organized larger QNs. More siblings are in various stages of incubation (e.g., QN-Control, QN-NSEEV and QN-ACES).
QN models were originally coded in commercial simulation software but are now being implemented in Python with easy-to-use interfaces. Model “Users” do not need to know Python. “Developers” can delve into and modify the codes.
Students/researchers/practitioners of any level of expertise can benefit from this workshop: learn how to use the QN methods/software (“get a feel for it”), incorporate it in dissertation/theoretical/applied research or product evaluation, and/or extend/challenge the models. A laptop is optional for hearing the lectures and seeing the demos but needed for accessing electronic documents.
Presenter
Arthur F. Thurnau Professor
Event Type
Workshop
TimeMonday, September 9th9am - 4:30pm MST
LocationFLW Salon G