Optimal design of twenty-five bar truss layout using a genetic algorithm

Well-known 25-bar truss is commonly optimized within a fixed topology and with area grouping. However, current paper attempts to investigate the optimal topology without prior assumption of the fixed topology. This paper employs a GA-based technique in which it departs from conventional GA where no binary strings is used to represent the possible solutions. The effect of relaxing the restriction of area grouping, and the variation of location of joints in the 25-bar truss are also investigated in finding the optimal weight of that given structure. For a given loading position, an optimal structure is found only with the requirement of at least three supports where the locations of the supports are also not fixed. Common optimization of this 25-bar truss not only fixed the number of supports to be four but also fixed the location of the supports. This paper deviates from this common practice where the optimization process should not be fixed with that requirement. It is found out that by using current algorithm, the optimal weight of 25-bar truss's topology is reduced. It proved that optimization without fixed topology can achieve less weight (i.e, approximately 30%) than with fixed topology by only knowing the loading joints.