Genetic algorithms in optimal design and detailing of reinforced concrete biaxial columns supported by a declarative approach for capacity checking

This paper introduces a new approach to optimal design and detailing of reinforced concrete biaxial columns using genetic algorithms (GAs). For a biaxial column with a given set of design requirements (section size, axial load and bending about both axes of the column), it is shown how GAs conduct a global search to identify the optimal reinforcement bar sizes and bar detailing arrangements. These satisfy maximum bending capacity about both axes of the column section and minimise the area of the reinforcements which leads to an economical design. In detailing reinforcement bar arrangements within the column section, the British Standard (BS8110) requirements are considered to ensure that both the ultimate limit state (ULS) and the buildability (ease of construction) requirements are satisfied. A declarative approach is used to check the exact bending capacities of the section about both axes of the column for the reinforcement bar detailing suggested by the GA, approved/modified and adopted by the designer. This is a final check which ensures that the column is designed and detailed properly and is capable of carrying the design loads safely. The human computer interaction (HCI) issue is an important focus of the current research in which the role of the designer as a decision maker and the role of the computer a decision support tool is maintained and enhanced.