selfadaptation

Self-adaptation is the capability of a system to autonomously alter its behavior, structure, or resource usage in response to changes in its environment or internal state, with minimal or no human input. It relies on built-in mechanisms that monitor, decide, and act to maintain or improve a system’s goals such as performance, reliability, or energy efficiency.

Most self-adaptive systems implement a loop that includes monitoring, analysis, planning, and execution, often referred to as the MAPE loop, and a knowledge base (MAPE-K). Monitoring collects metrics and events; analysis assesses deviations from goals or opportunities; planning selects adaptation actions; execution applies changes to the system. The knowledge base stores models, policies, constraints, and historical data to guide future decisions.

Self-adaptation is discussed across several domains. In biology, organisms exhibit self-adaptation through homeostatic regulation, phenotypic plasticity, and behavioral adjustments that help maintain viability. In software engineering and computing, self-adaptive systems autonomously adjust configuration, resource allocation, fault handling, or algorithms in response to workload shifts or failures. In robotics and cyber-physical systems, agents may reconfigure sensing, control parameters, or task plans to cope with changing tasks and environments.

Benefits include increased resilience, performance stability under varying conditions, and reduced need for human intervention. Challenges involve overhead, ensuring stability and safety, avoiding conflicting adaptations, and maintaining correct goals in dynamic contexts. Related concepts include autonomic computing, self-* properties, and adaptive control. Examples include cloud autoscaling, self-healing networks, database auto-tuning, and adaptive routing.