How one comes up with ideas for research: Remarks from personal experience

Two strategies for selecting research topics are possible: (1) Follow current publications and try to contribute; (2) Read good books (old and new ones alike), think, and create a new problem. The first strategy is safer; the second is far less assured but, perhaps, more satisfying and, potentially, productive. Some, more or less recent, examples of the latter are Shannon's information theory, Weiner's cybernetics, Bellman's dynamic programming, Kalman's /Letov's LQR/LQG, etc. Without comparing myself with these giants, in this talk I will describe how I arrived at a few ideas and what came out of them.