A Composite Algorithm for Mixed Integer Constrained Nonlinear Optimization

232 p.Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1980.A composite optimization algorithm applicable to mixed integer, constrained, nonlinear problems is developed in this research. One major component of the composite algorithm is a modified version of the nonlinear simplex method. Significant modifications are made to this algorithm including the incorporation of a unidimensional search procedure and the use of a new method to treat constraints. Additional features of the composite algorithm include new acceleration strategies, a new decomposition approach, and a discrete grid algorithm.The components of the composite algorithm are tested on problems primarily selected to represent engineering design optimization applications. The performance of the new methods is compared to some existing techniques. Examples of the application of combinations of the composite components are included. The results indicate that the new algorithms obtain superior solutions and in most cases are more efficient than existing techniques. The success of the algorithm on problems of engineering design optimization indicates a wide area of potential application.U of I OnlyRestricted to the U of I community idenfinitely during batch ingest of legacy ETD