Patterns of Behavior
Glenn Puchtel
Principal Software Engineer/Architect | .Net | Azure | DevOps | FinOps | IaaS | PaaS | SaaS | FaaS | AGI (Unconventional Intelligence) | Biocybernetics | Author (book, newsletter) | Prompt Engineer (LLMs)
Note: newsletters are progressive; each augment the previous. Reading them in chronological order is suggested for context. Previous: Cybernetic Waves, Next: Reaction Networks
"The significant problems we face cannot be solved at the same
level of thinking we were at when we created them."
~ Albert Einstein
Reality is the outcome of an infinite stream of cause & effect interactions. As things happen, it affects other things, causing things to happen. In truth, we are but machines acting in response to things that happen—behavior.
Over time, patterns of behavior have emerged dubbed archetypes. If we can recognize these patterns better, we can anticipate and react better—predictability. It is not surprising that our brain is a prediction engine; patterns appear everywhere; they fuel our brain.
Archetypes express predictable, cohesive systems. They provide clarity of intent whereby introducing false ideas is self-evident. That is, violations of a pattern are immediately and outwardly apparent.
For example, a thermostat exhibits the 'Balancing Process with Delay' archetype (shown later). It acts to achieve a goal by adjusting its behavior in response to delayed feedback. Introducing an external sensor to anticipate temperature fluctuations is outwardly wrong.
Fluctuations may cause premature actions or, worse, continuously wrong actions due to incorrect feedback from a failed sensor. Systems become less cohesive in this way; by adding improper dependencies—by violating their tenets.
Knowing the effect of archetypes suggests knowing the cause of behaviors. If we know that, we can reinforce good behavior and mitigate bad behavior. We can apply a (known) pattern to elicit desired results—the essence of software!
Archetypes
Cybernetics asserts success through archetypes. More so, through the causal relationships between them—feedback. Without quantifiable feedback and sufficient control, software is seldom successful. Worse yet, it is dependent on the skillset and timetable of those writing it. While programming principles, patterns, and practices are well known, feedback and control mechanisms are not.
Cybernetics is about feedback and control, and archetypes are models that express it. Archetypes do not propose solutions; rather, they offer strategies that are highly likely to succeed when combined—stories. If a story is clear, corrective actions to its variations are clear.
Strategies are either reinforcing or balancing structures that are affected by or affect another—positively or negatively. A reinforcing structure strengthens change in a particular direction (good or bad) through empathetic actions. In contrast, a balancing structure mitigates change through apathetic (goal-seeking) actions.
Each strategy has a suggestive title and is classified as:
- Parasitic: helpful to itself but harmful to another
-?Prey/Predator: mutually beneficial and harmful
- Reciprocal: mutually beneficial without harm
- Deterrent: harmful to another if another is harmful
- Altruistic: helpful to another, yet (possibly) harmful to itself
- Adaptive: multi-loop; adjusts control variables in response to environmental changes to achieve the desired goal—evolutionary.
Casual loops illustrate how structures reinforce or mitigate change; depicted by the following legend:
?-????????A negative correlation—both variables move in the opposite direction
?+???????A positive correlation—both variables move in the same direction.
?Bn??? A 'Balancing' feedback loop that seeks equilibrium; n = instance
?Rn??? A 'Reinforcing' feedback loop that amplifies change; n = instance
?//???????A ‘Delay’ in effect
Hint: To distinguish balancing from reinforcing archetypes, count the number of minus signs around the loop. An even number or zero indicates reinforcing. An odd number indicates balancing.
Cyborg: This tag, if present, expresses usage in Cyborg—the game
Process with Delay
In this regulatory archetype, the actual state interacts with the desired (fixed) state to produce a gap—the motive for corrective action (B1). The (delayed) action adds to the actual state, which deducts from the gap; when reduced to zero, no more corrective action occurs-thermostat.
Cyborg: Regulates body temperature. If too hot, sweating occurs; if too cold, shivering occurs.
Indecision
Whenever two goal-seeking structures interact, oscillations can occur. For example, B1 decreases price as supply increases. While B2 increases price as demand increases.
Naively, the structure should act as a Drifting Goals structure where the price stabilizes; instead, price variations cause a perpetual mismatch of supply and demand and oscillations.
Cyborg: Triggers smoothing algorithms in rules engine to minimize fluctuations by reducing the rate of change, delay, or relaxing the exit logic.
Escalation
In this integrative archetype, two balancing loops interact to create a single reinforcing loop. The response by one either increases or decreases this interaction.
Escalation occurs when the response by one increases the threat, corrective actions, or response of another. De-escalation occurs when the threat to or response lessens.
A non-cooperative pattern whereby both suppose that only one can prevail; a self-destructive loop unless one or both break the cycle.
Hint: considers competitive relationships among actors; understanding the perspective of one actor can impact the behavior of another—gameplay.
Cyborg: The interaction between the player (you) and the (Cyborg) character. Disruptions, either positive or negative, cause offsetting actions. Disturbances elicit cybernetic behavior by/from the Cyborg.
Drifting Goals
When the actions of one loop undermine the desired goal of the other. A gap between the goal and actual reality is resolved via corrective action (B1) or lowering the goal (B2).
A critical variation is that lowering the goal is instant, whereas corrective actions take patience.
领英推荐
Hint: Understand what indicators imply that goals are eroding? And what level of (dis)comfort affects expectations (goal)?
Cyborg: Persistent, abnormal vitals indicate that corrective actions are not effective or not occurring. Explore and invoke new trial-and-error measures, identify the root cause, settle for less, or accept a lesser outcome.
Fixes that Fail, Shifting the Burden & Addiction
An effective, short-term (symptomatic) solution creates side effects for a long-term (fundamental) solution, resulting in the need for continuous fixes.
The symptomatic solution (B2) seems better as there is no delay (B1). Sadly, it weakens the need for a fundamental solution—by lessening symptoms. The lack of a fundamental solution ensures problem symptoms return.
Hint: This structure can evolve into an Addiction structure—when the side effects overwhelm the initial problem symptom—when an agent becomes dependent on an external intervenor to suppress side effects—drugs.
Hint: 'Fixes that Fail' and 'Shifting the Burden' are fundamentally equivalent in cybernetics, as Cyborg makes no distinction.
Cyborg: high usage of a given ruleset indicates that algorithms are fixing symptoms, not problems—break the cycle by introducing entropy.
Accidental Adversaries
When two interacting actors (A & B) that should produce growth unwittingly limit the growth of each other.
A's activities toward B and B's activities toward A form a mutually beneficial outer reinforcing loop. The self-serving loops (R1 & R2) form one side of a balancing loop (B1 & B2), forming a mutually destructive loop (R3).
Cyborg: oxygen is exchanged for carbon dioxide (breathing); environmental changes inhibit carbon dioxide exchange for oxygen. Death ensues without intervention—assimilation of an artificial filter—toxicity.
Success to the Successful
Successful efforts consume a disproportionate number of resources, starving one or more competing efforts.
When allocations of a fixed amount of available resources favor one (R1), it lessens the amount available to another (R2). The greater the resources, the greater the success. The lesser the resources, the lesser the success. Eventually, one (R1) is starved of resources and fails.
Hint: a choice between preservation or extinction.
Cyborg: models conflict between and among two or more players in competition for resources.
Limits to Success
A once successfully growing system (R1) reaches limits; performance stalls regardless of how much time, effort, and resources are applied. At some point, constraints limit actions and are more significant than performance (B1).
Attractive Principle is a variation with constraints caused by two or more limits, each of a different form and typically dealt with in different ways, if at all.
Under Investment is a variation whereby an added loop justifies devesting the capacity to remove growth limits—that is, when what's required to remove limits is deemed too expensive. Sadly, the net effect, over time, can cause apathy and abandonment.
Hint: fosters the anticipation of constraints and limiting actions that impede long-term success.
Cyborg: indicates that assimilation of one or more devices is necessary for survival. At some point, no further actions are possible, and the player must choose a lesser effect, even if fatal. For example, if the temperature rises above the limits of available protective devices.
Tragedy of the Commons
Where multiple actors, behaving in their self-interest, escalate actions that deplete a shared resource and, over time, both exhaust the resource—exploitation.
It occurs when two or more reinforcing loops depend upon some common but limited resource. When the combined total activity depletes a shared resource, actors increase their activity to compensate, depleting it even faster.
Cyborg: general gameplay: total consumption or ruin of environmental resources till only complex cyborgs can survive—survival of the fittest. Conversely, feedback loops promote the growth of resources spent by cyborgs-conservation and contribution (by a player—you).
Conclusion
"Today's problems come from yesterday's solutions."
~ Peter Senge's – Fifth Discipline
Unintended effects threaten systems—behaviors cause effects—our behavior threatens systems. We must recognize and understand the complex causal chain of effects that influence behavior to mitigate threats.
Archetypes are patterns that depict and influence behavior. Cyborg—the game, features most, but not all of them—we're still exploring. Still, when thinking in cybernetics, we think of archetypes, determine if they are relevant, and structure our code and behavior accordingly.
We hypothesize that behavior is based on a finite set of rules governed by state. Voluntary or involuntary, behavior arises from a set of rules cultivated over time—evolution. In the absence of such knowledge, behavior is random.
Cyborg—the game, is a reactive rules engine defined by applied rules against transient or persistent conditions exhibiting behavior patterns-archetypes. Each has an inferred hierarchy, and each is explored in-depth in future newsletters:
Rules or a combination of them define conditions that elicit behavior. They are either deterministic or probabilistic—fixed or fluid. They range from simple validating, calculating (scores), categorizing, or transforming to complex decision-making (cognition) algorithms that invoke or discriminate between rules.
Discrimination requires evaluation. Studies show that dopamine signals in our brain help to evaluate between good and bad—the dopaminergic system (aka., the reward system).
Hint: In Cyborg—the game, we evaluate waves of chemical signals to elicit behavior.
We intend to offer a (C#) .Net set of integrated control structures fitting for gaming, enterprise, or cloud-based software warmly melded as C#yber.Net{ics}
We invite you to join us in constructive debates on our (LinkedIn) group as we continue a journey through cybernetic principles, patterns, and practices: