Although high-performance computing devices and general-purpose optimization software continue to be upgraded, optimization techniques are rarely utilized in the practical design of building structures. One reason is that optimization algorithms tend to be complicated and difficult to understand. This study proposes a section determination system for steel structures using a simple algorithm.
The results obtained by this study are as follows:
1. The optimization problem that minimizes steel weight was formulated by targeting the allowable stress design method of general steel structures, under the restricted condition of utility factor. Furthermore, in order to solve this problem easily, a simple algorithm consisting of decremental line search and automatic member grouping was proposed and the section determination system, combined with structure analysis, was built, analyzing the utility factor distribution and referring to the taboo list.
2. Termination criteria were proposed to easily reflect the design intent of the structural designer in the results of the proposed determination system. The system user can employ three parameters as termination criteria namely, the expected reduction ratio by member grouping, average utility factor of all section members, and number of groups.
3. Using a sample model of steel structures, the influence of each termination criteria on the design results was studied, and the range of each parameter to obtain practical design results was identified. Furthermore, in order to verify the validity of the proposed system, the obtained results were compared with those of structural designers, and the practicality of the proposed system was confirmed.
However, in the proposed system, only the utility factor for allowable stress method is considered as a constraint condition, hence other elements that are usually required for structural design such as eccentricity ratio and story drift angle are not considered. In order to develop further this practical system, it is necessary to take into account the previously omitted requirements. In addition, it is worth to include other requirements that seem difficult to deal with numerically such as workability.
