TRUSS BAR Problems Benchmark Suite: New Advanced Version

Next Generation and High Complexity TRUSS BAR Problems Benchmark Suite Source Codes and Competitive Meta-heuristic Search Algorithms

Highlights

• •Feasible solutions for nine different types of truss problems (TPs) were defined in this study for the first time.
• •LSHADE-CnEpSin, LSHADE-Epsin, LSHADE, SHADE, GSK, BES, FDB-AGDE, MPA and MFLA were applied for the first time for TPs.
• •The computational complexity and stability of algorithms were presented for the first time.
• •The results proposed in this article will ensure reusability and widespread impact for future research on TPs.

Abstract

Although they are among the most researched real world engineering design problems, it is encountered with significant problems in the optimization of structural truss bar problems (TPs). The main reasons for these problems are (i) the studies on the optimization of TPs being carried out with different experimental settings, (ii) the competitor algorithms used in the optimization process being insufficient, and (iii) the stability analysis and computational complexity information of the algorithms are not investigated. It is designed a simulation environment with defined standards and a benchmarking suite consisting of nine TPs of three different types in order to eliminate these problems. It is found optimum solutions for all problems in the benchmark suite and it is introduced feasible solutions for the first time. According to the statistical analysis results, among the seventy-seven competing algorithms, the best ten algorithms showed competitive performance in the optimization of TPs are LSHADE-EpSin, LSHADE-CnEpSin, SHADE, LSHADE, GSK, FDB-AGDE, MPA, LRFDB-COA and BES. According to the results of the stability analysis carried out on feasible solutions, it has been determined the algorithms with the best success rate for planar type; LSHADE, LSHADE-CnEpSin, LSHADE-EpSin and GSK with 100% success rate, LSHADE-CnEpSin, LSHADE-EpSin, LSHADE and SHADE for Space type with 100% success rate, and LSHADE-EpSin with 69% success rate for frequency constrained type. When the performances of the algorithms are evaluated regardless of the problem type, LSHADE-EpSin has been the most stable algorithm with an overall success rate of 90% on the TPs.