How did AI research in problem solving begin? Early research in AI often concentrated on games and puzzles. The study of games and puzzles stemmed from the belief that games and puzzles required intelligent behaviour for their performance. It was believed that AI should search for domain independent factors in problem solving - the so-called weak problem solving methods.
It eventually turned out that applying AI in these contexts is relatively easy compared with some of the other hard problems in understanding intelligence. For example, how babies learn their first fifty words, how to juggle five balls in the air, how to fluently translate from French to English.
What were the basic building blocks of problem solving programs? Initially, searching and sorting algorithms together with ``heuristics''. Note that an algorithm is taken here to be a finite sequence of steps guaranteed to perform some task while a heuristic is a useful rule of thumb which is not necessarily guaranteed to work but `worth a try'.