The marriage of on‐line human decision with computer programs

As the logistics problems of current interest become more and more complex, it becomes apparent that heuristic methods hold more promise than exhaustive or algorithmic methods of solution. It is also becoming clearer that heuristics, per se, are useless in approaching large combinatorial problems (via pattern recognition, for example), unless some general techniques are developed to draw out the specific “best” heuristics required for each problem. It is a fact that humans are, at present, far more efficient in developing and applying such heuristics than a machine. It is therefore of considerable value to explore how the best features of both human and machine decision can be married. The general problem requires the development of common meaningful languages, specialized input‐output display and entry facilities, and a means of time‐sharing the computer, in order to efficiently match the machine‐vs the human‐decision time.