Since dicts are ordered in modern Python, can we just use dict keys? I don’t know if there are any performance implications here.

Since dicts are ordered in modern Python, can we just use dict keys? I don’t know if there are any performance implications here.

Which is exactly what I did :)

@tpike3, @EwoutH, For the memory leak bugfix to pass, this PR needs to be merged first.

@quaquel Thanks for crushing on this to fix the issue.

Generally speaking, the difference between remove_constituting_agents and agent.remove is remove consituting agents removes the agent from the meta agent (e.g., they leave the alliance) but the agent is still a part of the model, agent.remove would remove them from the model entirely if I am not mistaken.

No, model.deregister_agent is what actually removes an agent from the model. Agent.remove() calls this. And, if you have a CellAgent or ContinuousSpaceAgent, agent.remove() also removes the agent from the space.

No, model.deregister_agent is what actually removes an agent from the model. Agent.remove() calls this. And, if you have a CellAgent or ContinuousSpaceAgent, agent.remove() also removes the agent from the space.

Hmmm I still thing there is still a bug in meta-agents/ the logic needs to be more transparent. This is currently not how it is written and was not touched by this PR. My goal with this is to write down my understanding until i/someone else has time to address. Document for the group.

User goal - remove agent from meta-agent but keep it in the model:
remove_constituting_agents should only remove form the meta-agent not the model

• line 384 in function remove_constituting_agents deregisters agent which will remove from the model not remove it from the meta-agent.
• Needs to be updated to just remove it from the meta-agent
• Needs to validate that when agent removed from model, it is also removed from meta-agent