Novel evolutionary algorithm with set representation scheme for truss design

Presented in this paper is a novel scheme of representation for truss geometry. Trusses are represented as a set of elements having a collection of properties (e.g. cross-sectional area, type of material). These sets can be of varying cardinality representing truss structures with different numbers of elements and hence distinctly different topologies. A recombination operator to handle a set representation that can generate offspring topologies that can be different from the parents is also proposed. Depending on the physical problem being solved, one can introduce specific operators which will aid the optimization process. One such mutation operator is used for the truss design to reduce the number of elements, hence finding the smallest feasible topology for the truss structure. Another mutation operator perturbs the properties of the elements using the Gaussian mutation.