Part of the project Decentralized Reality I wrote 2 scripts that visualize how distributed problems solve itself in a step by step procedure. These 2 problems explore both how agents in distributed systems can reach a consensus under imperfect communication.
----------------------------------
1.Byzantine Fault Scenario
-Red and green are generated randomly.
-When Red reaches consensus they attack.
-When Green reaches consensus they retreat.
-When they reach No Consensus they don't act.
----------------------------------
Each amazon knows that there are 17 unfaithful couples but does not know if she is one of them. She only sees the unfaithfulness of other couples. When the Amazon is certain of her betrayel, she kills her partner.
----------------------------------
Recursive Confirmation:
The absence of action for 17 days provides each amazon with a recursive confirmation of the number of unfaithful couples.
Each amazon uses the delayed action of others to deduce the total count accurately.
----------------------------------
Synchronous Deduction:
By the 17th day, the inaction on previous days confirms that the total number of unfaithful couples is indeed 17.
Each amazon has enough information to be certain of her own status and that of others.
-Green turns recursively into red in 17 steps.
-The unfaitfull partner (grey) disappears on step 17.