Agent-based modelling of socio-ecological systems: Models, projects and ontologies

Abstract

Socio-Ecological Systems (SESs) are the systems in which our everyday lives are embedded, so understanding them is important. The complex properties of such systems make modelling an indispensable tool for their description and analysis. Human actors play a pivotal role in SESs, but their interactions with each other and their environment are often underrepresented in SES modelling. We argue that more attention should be given to social aspects in models of SESs, but this entails additional kinds of complexity. Modelling choices need to be as transparent as possible, and to be based on analysis of the purposes and limitations of modelling. We recommend thinking in terms of modelling projects rather than single models. Such a project may involve multiple models adopting di↵erent modelling methods. We argue that agent-based models (ABMs) are an essential tool in an SES modelling project, but their expressivity, which is their major advantage, also produces problems with model transparency and validation. We propose the use of formal ontologies to make the structure and meaning of models as explicit as possible, facilitating model design, implementation, assessment, comparison and extension.

Energy dynamics for systemic configurations of virus-host co-evolution

“…There are known advantages of in silico modelling the action of therapeutic agents on known diseases through agent-based modelling 74 . However, the literature evidenced some intrinsic limitations on the choice of parameters, like the size of investigated populations 75 , while major problems are related to model validation 75,76 , also requiring to supplement the models with adequate formal ontologies 77 . Thanks to its abstract nature, stock-flow description can be used in a wide range of different fields, realizing the conceptual bridge that connects the language of biological systems to that of ecology.…”

Opportunities for agent‐based modelling in human dimensions of fisheries

“…In human dimensions of fisheries—and in the larger field of social‐ecological systems (SES)—the use of ABMs is becoming increasingly common within the scientific community, but uptake is still relatively limited in management settings (Gotts et al, ; Nielsen et al, ; Schulze, Müller, Groeneveld, & Grimm, ). (There have also been several ABM studies of fish and aquatic ecosystems, as noted above, but these are beyond the scope of our discussion here.)…”

“…ABMs have attracted scientific interest in these fields because groups of human agents interacting with each other and their environments have many features that ABMs are thought to be well suited to studying (Gotts et al, ). Specifically, (a) there are complex interactions between agents that influence the individual behaviour of each agent, and can produce complex aggregate behaviour patterns; (b) there is important heterogeneity among the agents in their decision processes or interactions with each other and their natural/social environments; (c) the agents follow decision rules that do not lend themselves well to analytical aggregation or abstraction (Axtell, ; Bonabeau, ; Conte & Paolucci, ; Gotts et al, ; Grimm, ; Heckbert et al, ; Schulze et al, ). Because ABMs are rooted in agent‐level decision‐making processes, they are flexible in complexity and scale, and provide a natural description of the system (Axtell, ; Bonabeau, ; Grimm, ).…”

“…Despite such advances, ABMs are still rarely used in fisheries management, especially in tactical situations (Nielsen et al, ). Reasons for this lack of uptake include the facts that ABMs are often complex, not analytically tractable and difficult to empirically validate—especially using statistical metrics of fit common to other types of models (Bousquet & Le Page, ; Gotts et al, ; Nielsen et al, ; Schulze et al, ). For these reasons, it can be challenging to communicate to managers how an ABM’s assumptions lead to its conclusions, and how to evaluate an ABM’s trustworthiness.…”

“…The experience of EBFM also illustrates the utility of diverse portfolios of modelling tools, spanning varying levels of complexity, to address both strategic (“big picture, direction‐setting and contextual”) and tactical (“focused on management actions on short timescales”) problems (Plagányi et al, ). The scientific and management‐related challenges in the human dimensions of fisheries demand a similarly diverse portfolio of modelling tools (Gotts et al, ).…”