Plenary lecture II: Stochastic resonance - An optimization means for manipulating uncertainty

Summary form only given, as follows. Conventional thinking has held that when a noise or some exonogeous disturbance is added to a physical system, generally most performance measures degrade accordingly. For certain classes of nonlinear systems, however, it is shown that there are some advantages in adding small amounts of noise in the sense that specific outputs may be enhanced to some degree. In the Physics literature, in recent times, there has been an explosion of research on a concept termed, “Stochastic Resonance.” This refers to optimizing some throughput characteristic of a nonlinear system via the abbbddition of a random signal. This methodology only works on nonlinear systems; it does not provide an equivalent benefit for a linear system. A number of examples are given from Physics, mathematics, and system theory to broaden the credence of the above statements. In this talk, data are presented from two real systems where a Stochastic Resonance design procedure was utilized to enhance certain key system response properties. The class of nonlinear systems that benefit from this approach are also defined. Finally, an overall design procedure is outlined on how