Multi-class teaching–learning-based optimization for truss design with frequency constraints

• A multi-class teaching–learning based optimization algorithm is proposed.
• The two-stage algorithm improves the exploration and exploitation of TLBO.
• Size and shape optimization under frequency constraints is addressed.
• Comparative studies illustrate the superiority of the proposed algorithm.

The primary objective of this study is to introduce a multi-class teaching–learning-based optimization (MC-TLBO) technique for structural optimization with frequency constraints. Teaching–learning-based optimization (TLBO) is based on a simple and efficient algorithm with no intrinsic parameters controlling its performance. The multi-class approach proposed here increases the initial exploration capability of the optimization process resulting in a more efficient search. MC-TLBO extends the concept of the education process from a single classroom to a school with multiple parallel classes. The MC-TLBO algorithm employs a two-stage procedure: in the first stage, parallel classes explore the search space; in the second stage, the best solutions of the first stage form a super class to be the initial population for a modified TLBO. In order to examine the efficiency of the proposed methodology, the MC-TLBO algorithm is applied on various benchmark truss optimization problems with frequency constraints and the designs results are compared to the results of both a modified TLBO algorithm and other optimization methods.