Abstract: A reference curve representing measured stress-strain data obtained in a material test of a specimen is received. FEA model is created to represent the specimen which is associated with numerical material properties defined by a formula having a set of adjustable coefficients. Multiple computed curves are obtained each defined with multiple nodes of computed stress-strain values by conducting a time-marching simulation of the material test using the FEA model with a set of unique coefficients. Respective curve difference measurement parameters are calculated between each computed curve and the reference curve using a similarity measure technique that includes trimming off excess end portion of each computed curve so as to match the reference curve. Optimal values of the adjustable coefficients are determined by estimating a minimum of the curve difference measurement parameters according to an optimization technique.

Abstract: A reference curve representing measured stress-strain data obtained in a material test of a specimen is received. FEA model is created to represent the specimen which is associated with numerical material properties defined by a formula having a set of adjustable coefficients. Multiple computed curves are obtained each defined with multiple nodes of computed stress-strain values by conducting a time-marching simulation of the material test using the FEA model with a set of unique coefficients. Respective curve difference measurement parameters are calculated between each computed curve and the reference curve using a similarity measure technique that includes trimming off excess end portion of each computed curve so as to match the reference curve. Optimal values of the adjustable coefficients are determined by estimating a minimum of the curve difference measurement parameters according to an optimization technique.

Abstract: Definition of a design space and an objective space for conducting multi-objective design optimization of a product is received in a computer system having a design optimization application module installed thereon. Design space is defined by design variables while objective space is defined by design objectives. First set of designs in the design space is selected. Each of the first set is evaluated in the objective space for non-dominance. Design space is partitioned into first and second regions using a multi-dimensional space division scheme (e.g., SVM). The first region is part of the design space containing all of the non-dominated design alternatives while the second region contains remaining of the design space. Second set of designs is selected within the first region. Each of the second set and existing non-dominated design alternatives are evaluated for non-dominance. Multi-objective optimization repeats the partition and evaluation until an end condition is reached.

Abstract: Definition of a design space and an objective space for conducting multi-objective design optimization of a product is received in a computer system having a design optimization application module installed thereon. Design space is defined by design variables while objective space is defined by design objectives. First set of designs in the design space is selected. Each of the first set is evaluated in the objective space for non-dominance. The design space is partitioned into first and second regions using a multi-dimensional space division scheme (e.g., SVM). The first region is part of the design space containing all of the non-dominated design alternatives while the second region contains remaining of the design space. Second set of designs is selected within the first region. Each of the second set and existing non-dominated design alternatives is evaluated for non-dominance. Multi-objective optimization repeats the partition and evaluation until an end condition is reached.

Abstract: Methods of conducting design optimization of a product using multiple metamodels are described. First and the second metamodels are configured with common kernel function. Kernel width parameter is the output or result of the first metamodel while the second metamodel requires a set of substantially similar kernel width parameters defined a priori. Further, the second metamodel is configured with an anisotropic kernel. First and second metamodels are trained in two stages. In the first stage, kernel width parameters are obtained by fitting known responses (obtained in numerical simulations) into the first metamodel with one or more prediction trends. Additional kernel width parameter set is derived by algebraically combining the obtained kernel width parameters.