Two-Stage Structural Optimization Design Based On Maximun Entropy

Based on maximum entropy two-stage optimization design theory, maximum entropy two-stage optimization design system including truss, beam, frame structures is founded in this thesis. Two-stage optimization design theory is extended to single-object and multi-objects non-linear mathematics planning. Examples show the method is concise and convergence rate is rapid. The contents are as follows:1. Maximum entropy two-stage optimization design of truss structures with continuous variables and discrete variables are further developed. Examples show convergence rate of optimization process is rapid whether they are continuous or discrete variables. Optimization design results are satisfaction.2. Modify the existing maximum entropy two-stage optimization design method of beam structures with continuous variables. Example of cross-beam shows convergence rate is rapid and continuous optimum solution approach global optimum.3. Found mathematics model of discrete variables for beam structures maximum entropy two-stage optimization design. Two-stage optimization design method of discrete variables for beam structures based on maximum entropy is present. Results show convergence rate is rapid and the times of convergence are irrelevant to the scale of the problems.4. Found mathematics model of continuous variables for frame structures maximum entropy two-stage optimization design method. Two-stage optimization design based on maximum entropy is present. Examples of portal frame, single -span two-storey frame, two-span six-storey frame are used to verify the validity of the method.5. Found mathematics model of discrete variables for frame structures maximum entropy two-stage optimization design method. Two-stage optimization design based on maximum entropy is present. To different constraint condition, examples of portal frame, single-span two-storey frame and two-span six-storey frame are used to verify the validity of the method.6. Present two-stage optimization method for single-object non-linear mathematics planning (SNLP) based on maximum entropy. Examples show the method is validity and convergence rate of optimization process is satisfaction whether they are continuous or discrete variables.7. Found mathematics model for multi-objects non-linear mathematics planning (MSNLP) based on maximum entropy. Present two-stage optimization method for multi-objects non-linear mathematics planning (MSNLP).