Human Performance Modeling and Simulation¶

Status: emerging
Last updated: 2026-05-31
Sources: 9781119636113.Ch26.Pdf, 9781119636113.Ch27.Pdf
Tags: [human-performance-modeling, mathematical-modeling, simulation, digital-human-modeling, queueing-network, cognitive-modeling, model-evaluation]

Summary¶

Human performance modeling represents human capabilities and limitations formally so that human-machine systems can be designed and evaluated before they are built (Wu & Liu, 2021; Paul, 2021). This article merges two chapters: one on mathematical modeling, which sets out four criteria for a good model and develops queueing-network approaches to single- and multi-task performance, and one on modeling and simulation of human systems, which spans biomechanical, cognitive, and physio-medical models. Together they show modeling as both a predictive design tool and a unifying framework for otherwise isolated concepts of human performance.

Body¶

Context¶

This article merges two handbook chapters that represent human capabilities formally so that systems can be designed and evaluated before they are built. Wu and Liu (2021) examine mathematical modeling in human–machine system design, setting out criteria for a good model and developing queueing-network approaches to single- and multi-task performance. Paul (2021) examines modeling and simulation of human systems, spanning biomechanical, cognitive, and physio-medical models. Each addresses a different layer of the same enterprise — predictive performance models and physical-physiological simulations of the body. Within this knowledge base the article is the modelling hub: Wu and Liu's queueing-network account formalises the processing stages described in Information Processing and the demands quantified in Mental Workload, while Paul's body simulations connect to the anatomical and posture work in Digital Human Modeling and 3D Anthropometry and to the timing of Selection And Control Of Action.

Key Points¶

Wu and Liu (2021) define four criteria for an ideal human-performance model: usefulness, robustness and generality, mechanism, and simplicity. Usefulness asks whether the model can improve a real system's performance, safety, or efficiency, distinguishing applied modeling from cognitive modeling that focuses on fundamental mechanisms. These criteria let designers compare mathematical models against artificial-intelligence models and other approaches on common terms (PDF p. 3, orig. p. 688).

Wu and Liu (2021) make queueing-network models their central technique for multitasking. The queueing-network (QN) framework incorporates multiple-resource models of divided attention and gives a computational account of both serial processing and concurrent execution. A three-node QN model represents task-selection strategy, server capacity, and queuing priority to handle a range of multitask situations, supported by experimental evidence. The same framework can model human-machine interaction in general, treating humans and machines as servers in a larger integrated network; like other task-analysis-based methods such as MHP/GOMS, QN-MHP grounds its predictions in decomposed task structure (PDF pp. 8–9, orig. pp. 693–694).

Paul (2021) situates modeling and simulation historically and biologically. Although modeling, simulation, and digital human modeling (DHM) are perceived as modern and dependent on powerful computers, the concept of a human system dates to Jastrzebowski's 1857 definition of ergonomics, in which a human is embedded in social and biological environments forming a system of systems. The physical human system divides into organ systems alongside psychological and social systems, setting up the multi-domain scope of human-systems modeling (PDF pp. 1–2, orig. pp. 704–705).

Paul (2021) organises models by domain and purpose: biomechanical modeling, physical-environmental and integrated modeling, cognitive and biologically inspired cognitive modeling, and medical and biologically inspired physio-medical systems modeling (PDF pp. 12–20, orig. pp. 715–723). Musculoskeletal human simulation (MSHS) is treated by its purpose, types, history, and applications in seat engineering, healthcare, and ergonomics generally (PDF pp. 20–21, orig. pp. 723–724).

Conclusion¶

The two chapters are complementary rather than competing. Wu and Liu (2021) supply formal, predictive mathematical models of performance, judged against usefulness, generality, mechanism, and simplicity, while Paul (2021) supplies physical and physiological simulations of the human body across biomechanical, cognitive, and physio-medical domains. They converge on the same purpose — design and evaluation before a system exists — but differ in level: Wu and Liu model the timing and structure of cognitive performance, whereas Paul models the body as a system of systems. Integrating the two into a single human-systems model remains open.

References¶

Jastrzebowski, W. (2000). An outline of ergonomics, or the science of work based upon the truths drawn from the Science of Nature, Part 1. Nature and Industry (Vol. 29, pp. 227–231). Republished by Central Institute for Labour Protection (2000). [Originally published 1857.] To be validated.

Paul, G.E. (2021) 'Modeling and Simulation of Human Systems', in Salvendy, G. & Karwowski, W. (eds.) Handbook of Human Factors and Ergonomics. 5th edn. Hoboken, NJ: John Wiley & Sons. paul2021modelingsimulation

Wu, C. & Liu, Y. (2021) 'Mathematical Modeling in Human–Machine System Design and Evaluation', in Salvendy, G. & Karwowski, W. (eds.) Handbook of Human Factors and Ergonomics. 5th edn. Hoboken, NJ: John Wiley & Sons. wu2021mathematicalmodeling

Open Questions¶

How do mathematical human-performance models compare with AI/machine-learning models on Wu and Liu's (2021) four criteria, especially mechanism and simplicity?
How can biomechanical, cognitive, and physio-medical simulations be integrated into a single human-systems model?