The prospects and limits of algorithms in simulating creative decision making

Although different types of algorithms and Turing-complete machines that underlie agent based models have opened up new vistas of scientific discovery to simulate decision-making by heterogeneous agents in artificial societies, we argue in this paper that there are significant limits of the algorithmic approach for simulating creative decision making by intelligent agents in rapidly shifting environments and social dynamics. In particular, we elaborate upon four specific limits of algorithms: (1) the decision rule and strategy space assignment problem; (2) the 'framing' or 'affordances' problem; (3) the problem of translation between higher level and base (Boolean) languages; and finally (4) the problem of choosing appropriate space-time scales for assigning decision rules to intelligent agents. We propose future development of non-algorithmic information processing systems, derived from genetic algorithms and learning classifier systems on the one hand and Neurosciences and Neurobiology on the other hand, to simulate creative decision making.
Turing Machines And The Advent Of The 'Age Of Algorithms' I believe that in about fifty years' time it will be possible to programme computers, with a storage capacity of [...]