At its core, in automatic control there are two interrelated emblematic problems, namely of stability and optimality of control. Classical control theory relates primarily to study of stability of closed loop systems, mainly to determine locus of stability curve and in particular the point where the closed loop system becomes uncontrollable. Optimality of control is confined to a narrow range of operation within the stability envelope. A certain combination of parameters such as unmeasured disturbance, rate of change of set point or process variation often cause the closed loop to deviate from its desired range and in the extremes induce oscillations leading to severe instability. 

​

For most part closed loop systems are designed to remain under "controlled stability" envelope and as for optimal performance takes a back seat. In practice, most control loops operate with "de-tuned" controller just so to avoid being unstable. 

​

In a PID control loop, the problems relating to stability and optimality of control are deeply entrenched. To gain stability, optimality is let go by "de-tuning" and to gain optimality (meaning faster response), stability loop is compromised. An oscillating loop with PID control is just trying to keep faster rate of response and loss of controllability out of each other. In practice, an oscillating loop is really not doing effective control and in effect always inefficient in terms of energy usage and/or product quality giveaway or violation.

​

This anomaly of reconciling the tradeoff between stability and optimality can best be understood by examining the following:

​

Why and how is that an open loop stable process becomes unstable when it is in closed loop?

​

​

AI-PID Tuning is predicated on the basic proposition that how not to make an open loop stable process becoming unstable in closed loop and how to make closed loop process to be more responsive than open loop at the same time every time. That is make closed loop system unbreakable and yet optimal no matter how little is known about the process and how quickly it changes.

​

After many many iterations of design studies, the founder of PIDAnalytix stumble across the truth of inherent stability and optimality. This led to formulation of AI-PID Tuning as audition herein.

​

At its core all control theory and practice is to achieve closed loop stability under varying conditions. That is, to avoid a control  loop becoming uncontrollable let alone perform optimally. Stability overrides performance. In this regard all control theory is all about scoping the locus of stability and avoidance of instability by various methods of analysis (in the authors' view paralysis) such as Bode Diagram, Nyquist theory and other. All of these classical methods have clearly not able to achieve both inherent stability and optimal loop performance equally satisfactorily. It is either about Stability or Optimality, seldom together.