Constraint first energy efficient indoor environmental control w 2027 Applie

Energy-efficient indoor control is often formulated as direct energy minimization, with comfort and indoor environmental quality (IEQ) represented by weighted objectives or soft penalties. Such formulations can report low energy use even when indoor conditions remain outside acceptable IEQ bounds. This paper proposes CKM-CF, a Constraint-first Knowledge-Memory Multi-agent Controller for qualified low-energy room-level operation. CKM-CF coordinates heterogeneous room devices through actuator-bounded device agents, deterministic conflict-aware bundle planning, and execution-derived knowledge-graph memory. Candidate bundles are selected by a constraint-first priority hierarchy: hard IEQ feasibility and task adequacy must be satisfied before energy is considered among qualified alternatives. The controller is evaluated in a simulated single-zone room-control environment, demonstrating effective management of energy consumption while adhering to strict indoor quality standards. Across multiple episodes, CKM-CF achieves a qualified performance rate significantly exceeding baseline methods, highlighting its potential for practical applications in smart building energy management.