IMPLEMENTING ANTIFRAGILES: SYSTEMS THAT GET BETTER UNDER CHANGE

摘要

Just as human bones get stronger when subjected to stress and tension, and rumors or riots intensify when someone tries to repress them, many things in life benefit from stress, disorder, volatility, and turmoil. Antifragiles are a category of systems that not only survive in such environments but actually get better. Current system engineering methods focus on survivability and stability, but how can antifragile systems be architected and implemented? The answer proposed in this paper is the use of the agent metaphor to enable a fuller treatment of systems' functionality. Instead of defining system functionality as fixed functions, system engineers define agents providing this functionality, but in addition can learn and improve in an agile environment. The paper presents a taxonomy of learning agents applicable in this context as well as the related utility functions required to enable the learning in an environment where the engineer may not even anticipate what the system shall learn in the changing future.