Speaker
Description
Softwarized service management increasingly relies on intent-based coordination, where feasibility is assessed against noisy operational telemetry rather than static thresholds. Crisp threshold validation is brittle under this noise, causing repeated accept-reject oscillations that destabilize automated closed-loop management. We propose a fuzzy satisfaction model that replaces binary feasibility with a graded, telemetry-derived score, integrated as a utility-shaping term in a conflict-graph selection problem while keeping hard policy and capacity constraints crisp. In a synthetic multi-KPI evaluation against a crisp baseline, the fuzzy formulation reduces selection instability and increases service-level utility, with the advantage confirmed under an externally defined validation metric that shares no parameter with either method. The approach remains compatible with standard optimization solvers and degrades gracefully to a lightweight heuristic for larger-scale deployments, supporting its use as a practical analytics-driven signal for trustworthy, closed-loop service and application management.