Simulation Overview - Concepedia

Principles

Definition

Simulation is a methodological approach employed across diverse academic and scientific fields to model and analyze the behavior of real-world or hypothetical systems and processes. It involves creating a representation, often computational, of the system under study, allowing researchers to perform experiments and explore various conditions or scenarios that may be impractical, costly, or impossible to replicate directly. This technique is significant for understanding complex dynamics, testing hypotheses, making predictions, and optimizing system design.

Ontological type

Core Methods

Applications

Modeling Frameworks

Key developments

Key Figures

Computational Simulation Foundations era

A. Rahman's 1964 molecular dynamics study showed that simulating particle trajectories could reproduce liquid-state behavior, establishing MD as a practical tool for physics-based simulation. Kirkpatrick, Gelatt, and Vecchi (1983) introduced simulated annealing, a probabilistic search strategy that dramatically improved optimization in large solution spaces. Car and Parrinello (1985) developed Car-Parrinello molecular dynamics, marrying electronic structure calculations with nuclear dynamics to enable more integrated simulations of materials. George E. Box and Norman Draper (1987) articulated surrogate modeling and uncertainty quantification through empirical model-building and response surfaces, guiding calibration and predictive simulation.

Scaling and Timescale Acceleration era

Donald B. Rubin [1] is a prominent statistician whose work in this era is associated with Harvard University [3] and University of California, San Francisco [4]. His contributions include the paper 'Inference from Iterative Simulation Using Multiple Sequences' [6], which advanced Monte Carlo methods and enabled scalable, multivariate inference in large-scale simulations. Andrew Gelman [2] is a statistician whose work in this era is associated with Harvard University [3] and University of Pennsylvania [5]. His contributions include the paper 'Inference from Iterative Simulation Using Multiple Sequences' [6], illustrating early development of iterative simulation and scalable Bayesian computation in this period.

Open Platforms and Digital Twins era

David A. Wood [1] is a prominent figure associated with Harvard University [3] and Stanford University [4] during this era. His contributions include the gem5 simulator [7], a modular, open-source architectural simulator enabling reproducible experiments and cross-disciplinary validation. Mark D. Hill [2] is a leading researcher linked with University of Wisconsin–Madison [5] and Duke University [6] in this period. The gem5 simulator [7] likewise exemplifies the era's open-platform spirit, providing a versatile, community-driven platform for validating architectures and exploring surrogate-driven models in digital twin-inspired experiments.