Reinforcement Learning: An Alternative Approach To Machine Intelligence

There are many unsolved problems that computers could solve if the appropriate software existed. Flight control systems for new, nonlinear aircraft, automated manufacturing systems, and sophisticated avionics systems all involve difficult, nonlinear control problems. Many of these problems are currently unsolvable, not because current computers are too slow or have too little memory, but simply because it is too difficult to calculate what the program should do. If a computer could learn to solve the problems through trial and error, that would be of great practical value. Attempts to solve this problem are generally known as artificial intelligence or machine intelligence. There have traditionally been two approaches, or paradigms, for creating useful machine intelligence. The first, the symbolic processing paradigm, assumes that we can tell the computer all the relevant facts in a situation, using a higher-level language than Ada or C. The computer is then expected to logically reason out what it should do, using a large number of IF-THEN type rules. For example, we might tell the computer all the facts about navigation and threat avoidance, and then expect it to logically deduce the best route for an aircraft to fly. Expert systems are the best example of this approach