BIONIC OPTIMIZATION OF THE CONTROL OF NONLINEAR DYNAMIC PROBLEMS

Abstract eng:
For nonlinear control problems the use of bionic optimization methods has been tested and qualified in many different fields [1, 2, 3]. The use of such optimization tools to derive the control system has the advantage of general applicability even in the case of very irregular responses and external influences to be handled. The fact that the derivation of the control parameters takes essentially more time than classical strategies like Laplace transformations seems to be prohibitive at a first glance. But as the control system might be used for a wide range of comparable tasks, the once found base solution has a large and often stable range of covered problems. So the relative time consuming set up time is well accounted for. Nevertheless the non-uniqueness of the solutions especially at hard nonlinear attacks inhibits the danger to select proposals that seem to be convincing at the first tests, but prove to be less powerful when entering real application scenarios. Therefor the decision making, which of the god solutions is really capable to cover the field of loading and other varying input has to be done with some care and experience. Taking into account all the critical remarks we expand the strategy to the question of robustness and reliability [4]. The well accepted ideas to perform robust and reliable optimization in a given space of scattering parameters and input data at control problems might be done by re-interpreting them as minimization problems. Response Surface type answers help to transform the high dimensional solution space into systems that might be dealt with in reasonable time. So a wide field of solutions and the influences of the ever present scatter might be handled without too much introduction of new tools but by application of qualified approaches to a new class of tasks. The control of cranes, which defines a well-known range of examples to the applicability of control problems is used to demonstrate the methods used and might help to apply the bionic methods to even more elaborated applications. For a set of questions we demonstrate the way to find a base solution. The base solution is expanded to handle the scatter of the input and system data. Using the following region of validity of the solution the whole range of expected transport histories is found within the range of applicability.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket