And-or graphs and theorem-proving graphs determine the same kind of search space and differ only in the direction of search: from axioms to goals, in the case of theorem-proving graphs, and in the opposite direction, from goals to axioms, in the case of and-or graphs. We investigate the construction of a single general algorithm which covers unidirectional search both for and-or graphs and for theorem-proving graphs, bidirectional search for path-finding problems and search for a simplest solution as well as search for any solution. Indeed, many different search spaces can represent the same original problem, as in the case of resolution systems where different refinements of the resolution rule determine different search spaces for a single problem of demonstrating the unsatisfiability of a given set of clauses. In the tree representation of theorem-proving graphs (used in Machine Intelligence 5 (Kowalski 1969)), identical problems Tare generated at different times, working forwards from the initial set of axioms {D, E, F, G}.

Continuing his exploration of the organization of complexity and the science of design, this new edition of Herbert Simon's classic work on artificial intelligence adds a chapter that sorts out the current themes and tools -- chaos, adaptive systems, genetic algorithms -- for analyzing complexity and complex systems. There are updates throughout the book as well. These take into account important advances in cognitive psychology and the science of design while confirming and extending the book's basic thesis: that a physical symbol system has the necessary and sufficient means for intelligent action. The chapter "Economic Reality" has also been revised to reflect a change in emphasis in Simon's thinking about the respective roles of organizations and markets in economic systems.

Merriman, J.H.H. | Wass, D.W.G.

Mr. J. H. H. Merriman was educated at King's College School, Wimbledon, and King's College, University of London. in 1935 and did Postgraduate Research at King's College London obtaining his M.Sc. We are, therefore, likely to see in the immediate future a movement away from the concept of single purpose automatic data processing installations to installations or systems of installations which, in the first placeT, will be multi--purpose and, in due course, integrated. If we are, therefore, to imagine large complex multipurpose integrated data processing system, we must imagine them to be serviced, to an increasing extent, by separate installations which will analyse the operations of the integrated system, determine the most appropriate operating conditions and which will, to some extent, relieve the burden of programming by automatic access to inbuilt programming routines.

These are (1) based on the use of conditional probabilities, (2) suggested by the idea that biological learning is due to facilitation of synapses and (3) based on existing statistical theory dealing with the optimisation of operating conditions. Although the application of logical-type machines to process control involves formidable complexity, design principles are evolved here for a learning machine which deals with quantitative signal and depends for its operation on the computation of correlation coefficients. By such trial-and-error procedures the control functions can be made to approach their optimum forms. THE CONDITIONAL PROBABILITY COMPUTER The ideas of probability theory must obviously be involved in any empirical approach to process control, since the aim is to maximise the probability of the desired goal in the future.