To be able to fulfil complex tasks it is crucial that agents share their maps. Currently, we only make teams of agents in the same map since it is the only way we have found to schedule a meeting and connect blocks. However, in many situations the agents are taking too long to meet each other and share their maps.
The image below shows agents' maps situation on the step 230 in a match of 300 steps. As we can see, there are only two agents sharing maps. They are the only ones that are able to perform any task.
It brings us many problems:
We are losing many tasks that were presented before the step when we have at least a pair of agents
Since we are sharing too slowly, we are losing other tasks since this only pair is busy
We don't have more options to try to make better teams (e.g. try to find the two closest agents to perform a given task)
To be able to fulfil complex tasks it is crucial that agents share their maps. Currently, we only make teams of agents in the same map since it is the only way we have found to schedule a meeting and connect blocks. However, in many situations the agents are taking too long to meet each other and share their maps.
The image below shows agents' maps situation on the step 230 in a match of 300 steps. As we can see, there are only two agents sharing maps. They are the only ones that are able to perform any task.
It brings us many problems: