How does DAGKNIGHT's approach to consensus differ from PHANTOM's?

DAGKNIGHT replaces PHANTOM's fixed-parameter approach with one that discovers the right parameter automatically. PHANTOM requires knowing a value called k in advance — k roughly represents how many blocks can be created in parallel — and then solves a problem called 'Maximum k-cluster subDAG' using that preset number. DAGKNIGHT instead solves the 'Minimal k Majority Cluster subDAG' (MkMC) problem, which means it figures out the appropriate k on its own rather than assuming it ahead of time. For a beginner, this matters because it means DAGKNIGHT can adapt to real network conditions instead of being locked into an assumption that may not match what's actually happening on the network.