Truss optimization with frequency constraints using the medalist learning algorithm

Though the truss optimization with frequency constraints is very complicated due to its high nonlinear property, the importance and necessity of truss optimization in the real engineering design render it attractive for many researchers. This paper proposed a new heuristic algorithm inspired by the observation of the learning behavior of individuals in a group for the truss optimization with frequency constraints. The new swarm intelligent heuristic algorithm called the medalist learning algorithm has a concise implementation procedure composed of two key operations: identifying the medalists and conducting the individual learning. The learning period is divided into predefined learning stages. In each learning stage, a learning efficiency is generated using a learning function originated from the natural learning function. The key operations mentioned above are repeated in each learning stage. For an excellent performer called medalist, to conduct individual learning is to carry out self-learning trials; but for a common learner, to conduct individual learning is to learn from his/her past experi-ence and imitate the medalists. Six well-known truss optimization problems were employed to verify the effi-ciency of the new algorithm. The results obtained by the medalist learning algorithm were compared with those previously reported in literature. The new algorithm outperforms the others in terms of the best and average weights. The computational cost measured by the number of the function estimations of the structure for the new algorithm to obtain the best design usually is in the middle of the reported costs. In some cases, the new algo-rithm has the lowest worst weight and smallest number of function estimations.