[ simulai.org ]
00 The Institute
VOL. LIV · ISSUE 03 · MMXXVI
[ Founded 1971 · Computational Simulation Studies ]

The Institute for Computational Simulation Studies

A non-profit research foundation for the study of synthetic systems, emergent behaviour, and the mathematics of imagined worlds.

Est. 1971
Fellows 248
Publications 1,832
Chapters 17

For more than five decades the Institute has served as the quiet, independent home of scholarship in computational simulation. Its founding charter1 describes the subject modestly -- "the careful study of imagined systems and the mathematics by which they are brought into being" -- yet the work has grown to encompass agent-based modelling, probabilistic inference, cellular automata, climate emulation, and the theoretical underpinnings of what is now commonly called artificial intelligence.

This document, our annual address to the fellowship and to the public, is presented in the tradition of our printed journals2: numbered sections, marginal indexes, ornamental dividers, and the careful footnote. We invite you to read at a measured pace.

§ 02 Mission and Method

On the careful study of systems that do not, strictly speaking, exist.

The Institute's mission is to advance the understanding of simulation as a distinct intellectual discipline -- neither wholly mathematics, nor wholly engineering, nor wholly philosophy, but drawing upon each. A simulation3 is, in our working definition, any system whose behaviour is intended to correspond to the behaviour of another system, real or hypothetical, under a stated set of rules of correspondence.

Our method follows three commitments, held since the Institute's earliest days and reaffirmed at each decennial congress:

  1. I.

    Rigorous formalism.

    Every claim about a simulated system must be reducible, in principle, to a statement about the rules from which it was generated.

  2. II.

    Patient empiricism.

    A simulation is a thing to be observed as well as constructed; its emergent behaviour is rarely self-evident from its definition.

  3. III.

    Open scholarship.

    All findings, including negative results and null experiments, are published openly in the Institute's quarterly journal4.

A simulation teaches us nothing about the world it imitates; it teaches us, rather, about the rules by which we have chosen to imitate it.

— E. Marquand, inaugural address, 1971

§ 03 Fields of Study

The twelve chairs of the Institute, in alphabetical arrangement.

Agent-Based Systems CHAIR · 01
Decision Trees & Branching CHAIR · 02
Recursive Structures CHAIR · 03
Probabilistic Inference CHAIR · 04
Convergence Analysis CHAIR · 05
Cellular Automata CHAIR · 06
Climate Emulation CHAIR · 07
Layered Learning Systems CHAIR · 08
Finite State Systems CHAIR · 09
Control & Feedback CHAIR · 10
Monte Carlo Methods CHAIR · 11
Phase Space Dynamics CHAIR · 12

A full catalogue of fellowships, reading lists, and current seminars within each chair is maintained in the Institute's public registry.5

§ 04 A Brief History

Half a century of quiet scholarship, in six entries.

  1. 1971

    Foundation.

    The Institute is incorporated in a former municipal library, with eleven founding fellows drawn from mathematics, meteorology, and early computer science. Its charter is signed on the autumnal equinox.

  2. 1978

    First Journal.

    The inaugural issue of Proceedings in Simulation Studies is printed in an edition of four hundred, and mailed to subscribing university libraries.

  3. 1986

    Cellular Chair Endowed.

    Following Dr. M. Okonkwo's landmark monograph On the Life of Arbitrary Grids, the Chair in Cellular Automata is formally endowed and occupied.

  4. 1994

    International Congresses.

    The first triennial International Congress on Simulation Studies convenes in Uppsala, drawing delegates from nineteen nations. The tradition continues to this day.

  5. 2008

    Open Access Charter.

    The Institute adopts its open access charter, placing all past and future publications in the public domain without embargo. Printed journals continue to be produced for the archive.

  6. 2023

    The Marquand Laboratory.

    A new computational laboratory is dedicated in honour of the Institute's first director. It houses a modest but permanent staff, a long reading room, and a fine view of the river.

§ 05 On Correspondence

A meditation offered at the 2024 congress, reproduced here in its entirety.

The question that has occupied us, from our first congress to our most recent, is neither how to build better simulations nor how to make them faster. It is, more plainly, the question of correspondence -- how a simulated system comes to mean something about a system which it is not.

A simulation of weather is not the weather. A simulation of a population is not a population. A simulation of a mind6 is, most certainly, not a mind. And yet each of these artefacts, carefully constructed, may teach us something true about its referent -- provided we hold clearly in view the rules by which the correspondence has been arranged.

This is the quiet ethical labour of our discipline: to keep the rules of correspondence in plain sight, and to resist, always, the temptation to forget that we have drawn them ourselves.

— H. Vandermeer, Director XIV · III · MMXXIV

§ 06 Recent Publications

A selection from the current quarter, freely available from our archive.

  • PSS 54.03.001

    On the Entropy of Arbitrarily Initialised Grids

    M. Okonkwo, J. Alvarez-Pérez

    We revisit the classical question of entropy growth in deterministic grid systems, proposing a refined measure that accounts for the observer's prior knowledge of the initial condition.

    CELLULAR AUTOMATA
  • PSS 54.03.014

    Patience and the Markov Chain: A Fieldworker's Account

    R. Nakamura

    An essay, rather than a paper, on the discipline required to observe slow-mixing chains without prematurely declaring their convergence.

    PROBABILISTIC
  • PSS 54.03.027

    Agents Without Objectives: Notes on Curiosity-Driven Traversal

    L. Fogarty, H. Vandermeer

    We describe a family of agent architectures whose behaviour is generated not by optimisation but by sustained interest in the unfamiliar regions of their state space.

    AGENT-BASED
  • PSS 54.03.041

    A Note on the Teaching of Simulation

    E. Harriman

    Reflections on a thirty-year seminar, with particular attention to the pedagogical value of failed simulations and the careful examination of their failure.

    PEDAGOGY

§ 07 Colophon

On the making, distribution, and correspondence of this document.

Typography

Set in Lora for display and Noto Sans for body, with DM Mono used for numerals and reference codes. Typography reviewed by the Institute's in-house compositor.

Correspondence

The Registrar of the Institute receives written correspondence by post to the Marquand Laboratory, or by electronic message at registrar at simulai.org.

Fellowship

Applications for junior, visiting, and honorary fellowships are considered annually by the Governing Council. A reading list and application guide is furnished upon request.

Archive

The full archive of our Proceedings, dating to 1978, is maintained in print within the Marquand Reading Room and in digital facsimile within the Institute's public registry.