MFRL - Khan BMS Battlefield Management System
Working notes on MFRL (Mean-Field Reinforcement Learning): ai & multi-agent context, design trade-offs, and where it fits in the Arban–Tumen hierarchy.
MFRL is a cost-curve question disguised as a technical one. If the per-node integration cost does not collapse, the standard does not matter.
Khan BMS doesn't ship MFRL as a checkbox. It ships it as the boundary between human authority and machine execution — signed at issue, verified at receipt, and replayable for any after-action review the JAG cares to run.
MFRL is anchored at the Arban — ten nodes under one tactical leader. Small enough to reason about by hand, large enough to absorb the loss of a node without re-planning. Authority for MFRL is bounded at this tier; nothing the Arban does can poison its parent.
MFRL is one of perhaps a dozen primitives that decide whether a modern force can fight through denial. Khan BMS is built on the premise that all of them deserve the same treatment.