Goals in Adaptive Systems
In the realm of adaptive systems, goals are a central concept around which different theorists have built their models. Three notable models that incorporate goals are Ashby's Law of Requisite Variety, Arthur Koestler's concept of bisociation, and the definition of intelligence offered by Marcus Hutter. This article explores the function of goals within each of these models and discusses how the principle of compounding can enhance the effectiveness of goal achievement.
Ashby's Law of Requisite Variety
W. Ross Ashby's Law of Requisite Variety is a foundational principle in cybernetics and control theory. It stipulates that for a system to be stable, the number of states that its control mechanism can attain must be greater than or equal to the number of states in the system being controlled. In this context, goals function as the desired states that the control system seeks to maintain or achieve. The variety in the control system must be sufficient to address the diverse challenges presented by the environment to meet these goals effectively. By applying the principle of compounding, a system can incrementally adjust its variety, learning and adapting from past experiences, thereby increasing its ability to swiftly and efficiently reach its goals with less energy expenditure over time.
Arthur Koestler's Concept of Bisociation
Arthur Koestler introduced the concept of bisociation to explain how creativity emerges from the intersection of two unrelated frames of reference. Goals in Koestler's model function as desired outcomes of creative processes. Bisociation allows for novel solutions by creating connections between disparate ideas. Through compounding, a frequent exposure to diverse domains and experiences can foster a richer substrate for bisociation, enhancing the likelihood of discovering innovative pathways to goals.
Markus Hutter's Definition of Intelligence
Deeply rooted in the study of artificial intelligence, Markus Hutter has defined intelligence as the ability to achieve goals in a wide range of environments. In Hutter's model, goals are benchmarks against which the performance of an intelligent agent is measured. According to Hutter, an intelligent system should be capable of learning and adapting to work towards its goals efficiently. The principle of compounding, when applied to intelligence, suggests that each learning experience builds upon the previous one, leading to a more refined strategy for goal achievement. This amplifies an agent's ability to reach goals faster and with less energy as it navigates increasingly complex environments.
Conclusion
In all three models, goals are pivotal in guiding system behavior, whether it's maintaining stability, fostering creativity, or maximizing intelligent performance. The principle of compounding acts as a force multiplier, leveraging past experience to improve future outcomes. By understanding the role of goals and the power of compounding, developers and theorists can design systems that excel in realizing objectives with greater speed, efficiency, and resourcefulness.