Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

General collections demography model with multiple risks

Josep Grau-Bové, Miriam Andrews

TL;DR

The paper tackles the challenge of modelling long-term heritage-collection degradation that combines slow continuous deterioration with infrequent catastrophic events, amidst sparse damage-function data. It proposes an Agent-Based Model with Monte Carlo sampling that integrates damage-function degradation, ABC risk parameters, and Weibull-based timing of adverse events to generate lifetimes and quantify uncertainty, implemented in R. Key contributions include a detailed ABM framework, the use of ablation studies to quantify the impact of individual degradation pathways, and proposed extensions such as dynamic environmental inputs and scenario planning to broaden applicability. The approach offers a practical tool for comparing risk factors, propagating uncertainty to lifetime estimates, and guiding conservation decisions, with publicly available code to enable reuse and extension.

Abstract

This note presents an Agent-Based Model (ABM) with Monte Carlo sampling, designed to simulate the behaviour of a population of objects over time. The model incorporates damage functions with the risk parameters of the ABC framework to simulate adverse events. As a result, it combines continuous and probabilistic degradation. This hybrid approach allows us to study the emergent behavior of the system and explore the range of possible lifetimes of a collection. The main outcome of the model is the decay in condition of a collection as a consequence of all the combined degradation processes. The model is based on six hypotheses that are described for further testing. This paper presents a first attempt at an universal implementation of Collections Demography principles, with the hope that it will generate discussion and the identification of research gaps.

General collections demography model with multiple risks

TL;DR

The paper tackles the challenge of modelling long-term heritage-collection degradation that combines slow continuous deterioration with infrequent catastrophic events, amidst sparse damage-function data. It proposes an Agent-Based Model with Monte Carlo sampling that integrates damage-function degradation, ABC risk parameters, and Weibull-based timing of adverse events to generate lifetimes and quantify uncertainty, implemented in R. Key contributions include a detailed ABM framework, the use of ablation studies to quantify the impact of individual degradation pathways, and proposed extensions such as dynamic environmental inputs and scenario planning to broaden applicability. The approach offers a practical tool for comparing risk factors, propagating uncertainty to lifetime estimates, and guiding conservation decisions, with publicly available code to enable reuse and extension.

Abstract

This note presents an Agent-Based Model (ABM) with Monte Carlo sampling, designed to simulate the behaviour of a population of objects over time. The model incorporates damage functions with the risk parameters of the ABC framework to simulate adverse events. As a result, it combines continuous and probabilistic degradation. This hybrid approach allows us to study the emergent behavior of the system and explore the range of possible lifetimes of a collection. The main outcome of the model is the decay in condition of a collection as a consequence of all the combined degradation processes. The model is based on six hypotheses that are described for further testing. This paper presents a first attempt at an universal implementation of Collections Demography principles, with the hope that it will generate discussion and the identification of research gaps.
Paper Structure (16 sections, 1 equation, 6 figures, 3 tables)

This paper contains 16 sections, 1 equation, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. An operational definition of lifetimes
  3. The model
  4. Modelling degradation rates with damage functions
  5. Definition of adverse events
  6. Modelling time to failure with the Weibull distribution
  7. Monte Carlo sampling of adverse events
  8. Agent-Based Model Description
  9. Implementation in R
  10. Example results
  11. Extensions to the model
  12. Incorporate Diverse Degradation Functions
  13. Dynamic Environmental Conditions
  14. Agent-Based Model Enhancements
  15. Scenario Planning
  16. ...and 1 more sections

Figures (6)

  • Figure 1: Condition decays in an unknown and unpredictable way. But the crux of decision making is focused on a small time window at the start of the process.
  • Figure 2: Summary of the capabilities of damage functions depending on their level of development.
  • Figure 3: Probability density function of the Weibull distribution for different values of $k$ and $\lambda = 1$.
  • Figure 4: Histogram of condition during a simulation
  • Figure 5: Simulation example showing 10 runs. Some runs exhibit large impact risks while others show many low impact risks, yet overall behaviors are very similar.
  • ...and 1 more figures