Cable-stayed bridge optimization solution space exploration

In this work, we explore and optimise solutions for a controlled cable-stayed footbridge with a pre-determined length and width. The problem consists of designing a complete and cost-efficient bridge solution with geometry, cross-section, prestress and control design variables. The dynamic and static constraints create a deceptive landscape of the solution space, which, in turn, makes this problem hard to optimise via standard optimisation. We start with a conventional Genetic Approach and show that we are able to fulfill the structural requirements. Afterwards, we move to explore the variables and conditions that enable us to traverse in the solution space, optimising structural requirements while lowering the cost of solutions. The results present the sensitivities for all design variables for the baseline solution. We observed that most design variables could not be changed in any direction without compromising the structural constraints or cost. The results demonstrate that the optimisation of cable-stayed bridges requires finding a very delicate balance between multiple structural constraints and cost, which poses an interesting benchmark problem for the use of evolutionary algorithms in structural optimisation.