SOSIEL Algorithm

The open-source multi-agent SOSIEL (pronounced ˈsōSHəl and stands for Self-Organizing Social & Inductive Evolutionary Learning) algorithm simulates the spatio-temporal complexity of social contexts in which the heterogeneity of knowledge, the need for learning, and the potential for collective action play a significant role. It can simulate the cross-generational progression of one or a large number of boundedly rational agents, each of which is represented by a cognitive architecture that consists of theoretically grounded cognitive processes and agent-specific and empirically grounded knowledge. The agents can interact among themselves and/or with coupled natural and/or technical spatiotemporal systems, learn from their and each other’s experience, create new practices, and make decisions about taking and then take (potentially collective) actions.

Features

  • A multi-agent algorithm simulating social learning, collective action, cross-generational population dynamics, and the self-organization of multi-layered social network structures.

  • Agent cognition is represented with a general cognitive architecture that consists of a memory component, a learning component, and a decision-making component and that can be set to one of four cognitive levels.

  • ​Each agent can engage in anticipatory learning, goal prioritizing, counterfactual thinking, innovating, social learning, goal selecting, satisficing, signaling, and (potentially collective) action-taking.

  • Option to set agent cognition to one of four levels, with each level composed of a different combination of cognitive processes and representing an alternative approach to modeling agent behavior.

  • Agents can be empowered with place-based and/or hypothetical knowledge that is organized, updated, modified, and/or utilized by the cognitive architecture.

Requirements

To use the SOSIEL algorithm, you need to:

  • Join the SOSIEL Algorithm Users group. You MUST join this group before accessing SOSIEL installers.

  • Install a model that implements the SOSIEL algorithm (see below).

  • Check if the model includes the latest version of SOSIEL. If not, update the model with the latest version of SOSIEL (see download and installation instructions below).

  • Create input files for the model (see examples below).

Models running the SOSIEL algorithm and example input files

Release notes

Download

The installer for the latest version of the SOSIEL algorithm can be downloaded here.

Citation

Cite the following journal article when using SOSIEL on its own or in another model: Sotnik, G. (2018). The SOSIEL Platform: Knowledge-based, cognitive, and multi-agent. Biologically Inspired Cognitive Architectures, 26, 103-117.

Support

If you have a question or come across any issue or suspected bug when using the SOSIEL algorithm, email Garry Sotnik at: contact@sosiel.org.

Design

  • Garry Sotnik

Development team

  • Aleksandr Karifanov

  • Ivan Pizhenko

  • Garry Sotnik

  • Eugene Lobach