What Is a Mathematical Model?

A mathematical model is an abstract representation of a physical, biological, economic, or engineering system expressed in the language of mathematics, using variables, equations, functions, and logical relations to describe how the system behaves. Models translate qualitative understanding of a phenomenon into a form that supports quantitative analysis, prediction, and design. They range from simple algebraic expressions, such as Ohm's law relating voltage and current, to systems of partial differential equations governing fluid flow in a gas turbine.

Mathematical modeling draws on mechanics, statistics, numerical analysis, and domain-specific science, and it is practiced across all quantitative disciplines. The NIST Applied and Computational Mathematics Division conducts research in applied mathematics and scientific computing to develop modeling techniques with broad application in measurement science and engineering. The value of a model lies not in its complexity but in its ability to make testable predictions and to reveal the governing mechanisms that determine a system's response.

Types of Mathematical Models

Models are classified along several dimensions. Deterministic models produce a unique output for each set of inputs, while stochastic models incorporate random variables to represent uncertainty or variability in the system. Continuous models describe systems using differential or integral equations, while discrete models use difference equations or logical rules operating on countable states. Static models represent equilibrium conditions, whereas dynamic models track how states evolve over time. Lumped-parameter models treat a distributed system as a small number of interacting elements, simplifying analysis at the cost of spatial resolution; distributed models retain spatial variation by formulating partial differential equations over the system geometry. Each classification reflects a choice about which aspects of the real system are essential and which can be idealized away.

Model Formulation and Validation

Formulating a model requires identifying the variables that describe the system state, the parameters that characterize its properties, and the governing laws or empirical relationships that connect them. Conservation laws, constitutive relations, and kinetic rate equations often provide the structural skeleton, while experiments supply parameter values and boundary conditions. Once formulated, a model must be validated: its predictions are compared against independent measurements not used in calibration to assess whether the model captures the phenomena of interest within acceptable accuracy. Sensitivity analysis reveals which parameters most strongly influence the output, guiding where additional experimental data will most improve model fidelity. The NIST modeling and simulation research program addresses these validation challenges across manufacturing, materials, and measurement applications.

Simulation and Computational Methods

When analytical solutions are unavailable, which is the common case for nonlinear or high-dimensional models, numerical methods convert the mathematical formulation into a problem a computer can solve. Finite-element methods discretize continuous domains into meshes, reducing partial differential equations to large sparse linear systems. Finite-difference and finite-volume methods apply similar ideas on structured grids. Monte Carlo methods sample random variables to estimate statistical properties of stochastic models. The accuracy of a simulation depends on the numerical method's consistency and stability, the resolution of the discretization, and how faithfully the model captures the underlying physics. As discussed in NIST's mathematical models and complex disease research, even biologically complex systems can be represented with sufficient fidelity to guide experimental design and therapeutic strategy.

Applications

Mathematical models have applications in a wide range of disciplines, including:

• Electrical and electronic circuit design, where SPICE models simulate transient and steady-state behavior before fabrication
• Structural engineering, where finite-element models predict stress, deformation, and failure under load
• Climate science, where coupled atmosphere-ocean models simulate long-term temperature and precipitation patterns
• Control systems design, where transfer function and state-space models support stability analysis and controller synthesis
• Biomedical engineering, where pharmacokinetic models guide drug dosing and physiological models inform device design