Thanks for the reminder on the handshake formula. I made a few graphics to help describe the previous posts.Here’s what the objective space looks like for initial group sizes of 4 to 30 elders. The colors indicate the group size.
This is what happens to the feasible options when we demand <= 25% malicious node handling, and <=25% node loss /failure handling. Infeasible candidates are gray.
A 2D slice of the data showing a pareto (non-dominated) front in the upper left corner for maximal malicious node handling and minimal communication overhead. Infeasible candidates are gray.
A 2D slice of the data showing a pareto front in the upper left corner for maximal node loss/failure capability and minimal communication overhead.Infeasible candidates are gray.
If you increase the criteria to 33% malicious node handling and 25% loss ratio, then an initial group size of 12 elders is the feasible candidate with the least communication overhead. This is shown below in black. I find this to be a rather elegant candidate, since it also meets the sybil resistance criteria when 8 to 12 adults per elder.
