Coping With Problems Computers Can't Solve,

For over 40 years, the branch of computer science known as artificial intelligence has been unable to fulfill its promise of providing truly intelligent machines for general use. As early as the 1950s and 1960s, scientists developed computational models of intelligence then excitedly coded these models into the best computers of the day. At first the scientists were puzzled by the machines inability to produce reasoned output. This turned to frustration when they realized they had smashed into a thick brick wall which they had failed to forsee. This wall stopped them in their tracks and continues to do so today, decades later. It also has stymied scientists and engineers in other fields such as operations research the field concerned with determining efficient manufacturing and scheduling protocols, VLSI chip design and testing, and data base management to name a few. The brick wall exists because many combinatorial problems that are of fundamental importance to intelligent models, operations research, etc. are NP complete.