Module sp_npos_elections::balancing

Balancing algorithm implementation.

Given a committee A and an edge weight vector w, a balanced solution is one that

1. it maximizes the sum of member supports, i.e Argmax { sum(support(c)) }. for all c in A.
2. it minimizes the sum of supports squared, i.e Argmin { sum(support(c).pow(2)) } for all c in A.

See balance for more information.

Functions

balance

Balance the weight distribution of a given voters at most iterations times, or up until the point where the biggest difference created per iteration of all stakes is tolerance. If this is called with tolerance = 0, then exactly iterations rounds will be executed, except if no change has been made (difference = 0). tolerance and iterations are part of the BalancingConfig struct.