Skip to main content

Table 3 One of the eleven architectural styles for self-adaptive systems described in the catalog

From: Assessing the benefits of search-based approaches when designing self-adaptive systems: a controlled experiment

Style #2 – P(ID) Feedback control with system identification
Summary A separate component (controller) measures system output and acts accordingly to drive the system to the expected output (feedback)
Components System, controller, effector, sensor, transducer/QoS subsystem (optional, if output not directly delivered by the system itself)
Connectors (Remote) procedure call, event bus or data access
Data elements Reference value(s), input values, output values, transduced values (optional)
Topology Circular with one entry point: (reference input → controller → system → measured output [ → transducer/QoS subsystem] →...)
Variants #2.1: PID-SI with Precompensation - (PID-SI/PC)
  #2.2: PID-SI with Sensor Delay - (PID-SI/SD)
  #2.3: PID-SI with Filtering - (PID-SI/F)
Qualities yielded Reactive behavior; adaptation to unmodeled disturbances; no need for an accurate system model; can make stable an unstable system
Typical uses When a good enough system model is available, disturbance modeling is quite complicated, target system is unstable but linear or with identifiable linear operating regions
Cautions When disturbance spans over a wide range; when system is primary non-linear or have dynamics that are difficult to be modeled; when structural reconfiguration is needed
Example ...