Evolutionary Structural Optimization
The wonderness of design binerization
The optimization dependency of initial problem design variables promoting the utilization of binarization methods (i.e. the design domain is either zero or 1 during from the initialization phase, updating till final design) for concurrent multiscale topology optimization such as Evolutionary structural optimization (ESO), which has gained popularity, due to binarization of the initial design domain and the common doctrine of zeroing the central element value of the parodic structure to work as a trigger of inverse homogenization. Performing ESO with this prescribed setting will lead to uniformity of design as well as limiting the chances to fall into local minima. With covering almost all the design domains, ESO tries to reduce the volume to find the optimum solution as binary topology.
In general, ESO has been used with many extensions for multi-physics problems. The basic philosophy behind ESO is the direct removal of the so-called “inefficient material” which is leading to structure to form the optimal design. It is firstly introduced by Min and Steven. The cost function sensitivity is used to update the decision variables. Updating is depending on the element sensitivity number obtained by differentiating the objective function such that of solid elements and soft elements is equal to the elemental sensitivity and zero, respectively. In other words, redundant materials which identified by low sensitivity number will be eliminated. Because ESO is limited by only permitting redundant material removal from the structure, and those materials cannot be reintroduced later, an oversized initial design domain is essential to ensure that the final design is adequately represented. Bi-directional Evolutionary Structural Optimization (BESO) method was introduced to allow the creation rather than the removal of materials within the design domain, which increased the robustness of ESO and minimized the local extrema of the solution.