NetLogo User Community Models

(back to the NetLogo User Community Models)

Download If clicking does not initiate a download, try right clicking or control clicking and choosing "Save" or "Download". Try It in NetLogo Web

## WHAT IS IT?

This model investigates the success of cooperation when under ecological pressure, including when non-cooperative agents are also present in the system and take advantage of cooperative efforts for the survival of the group. In particular, two cognitive abilities are tested in such circumstances: memory and language, which are necessary to build and maintain a network of direct and indirect reciprocity, thus allowing for stable cooperation and better survival chances.

## HOW IT WORKS

100 agents are placed randomly in space with random initial energy between 1 and 100 (maximum possible energy). A population size of 50 to 100 individuals is considered based on estimates for the group size of hunters–gatherers societies.

Trees are randomly put in space, each with 5 fruits. Trees die when all fruit is collected. Ecological pressure is determined solely by the nature’s regrowth rate rr, that is, a new tree comes to existence whenever (ticks mod 2 x (2^3 - rr )) = 0.

Agents move randomly: one step per tick inside their perception field, defined by a 120º attentional angle and a 9 step attentional distance. Agents loose 1 unit of energy per step and when they loose all their energy they die. If a tree is detected in the agent’s perception field he moves towards the tree to collect one fruit. If more than one tree is detected he moves towards the closest tree. If other agents with energy greater than 33 are detected in his perception field and ego has energy less than 33, he sends an energy request to one of these agents. If the request is accepted, their energy is summed and divided by both. If the request is refused he may be able to memorize the other’s identity. If an energy transfer is successful, agents may also share information in memory, independently of who made the request. In that case, their memory after the interaction is the concatenation of their personal records (before the interaction) of the identities of the non-cooperative agents.

Agents can be cooperative (C) or selfish/non cooperative (NC). Cooperative agents are always available to share their energy (if greater than 33 units), unless they know their partner from previous interactions or "rumors" (conditional cooperation based on memory or memory and language). Selfish agents are never available to share their energy.

## HOW TO USE IT

Control the regrowth rate my manipulating rr. High, medium and low ecological pressure are defined by rr = 1, 2 and 3 respectively.

The initial number of cooperative agents in the system is controlled by C, the rest (100 - C) are not cooperative.

The number of agents an agent is able to memorize is controlled by memory-limit. (only relevant in experiments 2 to 4; See below)

Select experiment:

1. no memory or language.
2. cooperative (C) have memory.
3. cooperative have memory and language.
4. non-cooperative (NC) also have memory and language.

setup and go.

## THINGS TO NOTICE

How do you compare the impact of different rr values on global survival chances?

How is the success of cooperation changed by the initial number of cooperative agents in the system?

How to interpret the effect of memory and language on the success of cooperation?

## THINGS TO TRY

Try comparing the success of cooperation across experiments for different values of C and rr.

## EXTENDING THE MODEL

This model was developed specifically to test the importance of reciprocity and underlying cognitive mechanisms for stable cooperation under ecological adversity. As such, all parameters were set empirically such that, on one hand, survival chances are affected by ecological pressure, and on the other, promoted by an interaction frequency that is enough to allow many reciprocity opportunities. For that reason parameters where set implicitly in the code. But it could be interestign to manipulate all parameters parameters explicitly and try for example different proportions of energy sharing.

Ideas:

1. Create obstacles and let the agents follow the walls by avoiding them.
2. Allow the agents to reconsider their strategy based on their previous experience. Things like gratitute and other complex social emotions and moral rules may be interesting to investigate in this type of setting.
3. Allow agents to reproduce.

## RELATED MODELS

In NetLogo Modeling Commons:

Cooperation
Altruism
Reciprocal Altruism in Vampire Bats

## CREDITS AND REFERENCES

This model was originaly inspired by:

Zibetti, E., Carrignon, S., & Bredeche, N. (2016). ACACIA-ES: an agent-based modeling and simulation tool for investigating social behaviors in resource-limited two-dimensional environments. Mind & Society, 15(1), 83-104.

The present model was coded by David N. Sousa (davidnsousa@gmail.com) with the collaboration of Luis Correia and Leonel Garcia-Marques at Universidade de Lisboa. The model was developed in the context of a master's thesis by David N. Sousa: http://hdl.handle.net/10451/32768

Feel free to contact us with comments or questions.
July 25, 2018, Lisboa, Portugal.