A Design Methodology for a CSP-Based Optimization Approach

Design optimization is a common practice in industry. Different
mathematical algorithms have been developing to support the optimization in
engineering design. The integration between optimization methods and simulations
is an interesting issue in engineering design. A typical optimization workflow can
include simulation steps; however, the Virtual Prototyping analysis is more timeconsuming
than analytical calculations. The study of Constraints Satisfaction
Problems is a mathematical topic which can be applied for solving engineering
issues in design. The strength of this approach is the velocity on searching the
satisfied solutions. This paper proposes a design methodology which considers the
use of CSP models and calculation tools to optimize the sizing of columns and
beams in the design of a steel structure. The calculation tools regard analytical
models and numerical analysis. While the analytical approach regards the
computing of cost and weight, the numerical analysis is used to verify and check
the engineering performance in terms of deformation and stress state. A
customized application, based on MiniZinc platform, has been developed and
proposed to solve the CSP model for a test case steel structure. The CSP problem
has been limited to the calculation of analytical constraints such as cost and weight.
Finally, the resultant set of solutions has been evaluated using numerical solution
to complete the optimization analysis.