Evolutionary Black-Box Topology Optimization: Challenges and Promises

Abstract

Black-box topology optimization (BBTO) uses evolutionary algorithms and other soft computing techniques to generate near-optimal topologies of mechanical structures. Although evolutionary algorithms are widely used to compensate the limited applicability of conventional gradient optimization techniques, methods based on BBTO have been criticized due to numerous drawbacks. In this paper, we discuss topology optimization as a black-box optimization problem. We review the main BBTO methods, discuss their challenges and present approaches to relax them. Dealing with those challenges effectively can lead to wider applicability of topology optimization, as well as the ability to tackle industrial, highly-constrained, nonlinear, many-objective and multimodal problems. Consequently, future research in this area may open the door for innovating new applications in science and engineering that may go beyond solving classical optimization problems of structural engineering. Furthermore, algorithms designed for BBTO can be added to existing software toolboxes and packages of topology optimization.

CarHoods10k: An Industry-Grade Data Set for Representation Learning and Design Optimization in Engineering Applications

“…Nevertheless, heuristic approaches are not able to address specific objectives and constraints directly, which significantly limits their applicability. To address this problem, a field of black-box topology optimization (BBTO) emerged [36], offering a possibility to optimize any quantifiable objective functions with use of standard non-gradient optimization techniques such as evolutionary algorithms (EAs) or Bayesian optimization (BO). Here, also alternative, natureinspired meta-heuristics proposed in the recent years (e.g., [37], [38]) could be potentially used, as demonstrated in [39].…”

From Bioinspiration to Computer Generation: Developments in Autonomous Soft Robot Design

“…This class of solver comprises numerous heuristic and stochastic search algorithms, several of which are inspired by natural processes, such as particle swarm optimization, simulated annealing, and ant colony optimization. [ 95 ] They apply metaheuristic rules to search large and nonlinear spaces, enabling them to avoid local minima and potentially reach a global optimum in nonconvex problems, without problem‐specific knowledge. These heuristic methods explore a significantly larger region of the design space than greedy gradient‐based methods, allowing fitter candidates to be found.…”

“…In addition, stochastic nongradient methods also suffer from connectivity problems, where infeasible, discontinuous regions of elements occur within the design space. [ 95,96 ] In the context of soft robots, these manifest as open chambers in pneumatic devices or disconnected parts.…”

“…Morphogenetic methods based on DNA encoding have also been proposed to reduce the complexity of soft robot representation while incorporating greater detail. [ 51 ] For a detailed review of encodings used in topology and soft robotic optimization, see Guirguis et al [ 95 ] A summary of evolutionary soft robot optimizations is shown in Table 2 .…”

Part consolidation for additive manufacturing: A multilayered topology optimization approach

“…7 While gradient-based optimization techniques are most often used in TO, black box TO approaches may become an effective alternative for solving multiobjective, nonlinear optimization problems. 8 AM is an ideal candidate for fabricating topology-optimized designs as the resulting complex geometry can be printed with minimal interpretation, eliminating the associated reduction in performance. There has therefore been significant research focused on integrating TO algorithms with the AM driving factors for cost and performance, known as design for AM.…”