A novel chaotic Henry gas solubility optimization algorithm for solving real-world engineering problems

The paper proposes a novel metaheuristic based on integrating chaotic maps into a Henry gas solubility optimization algorithm (HGSO). The new algorithm is named chaotic Henry gas solubility optimization (CHGSO). The hybridization is aimed at enhancement of the convergence rate of the original Henry gas solubility optimizer for solving real-life engineering optimization problems. This hybridization provides a problem-independent optimization algorithm. The CHGSO performance is evaluated using various conventional constrained optimization problems, e.g., a welded beam problem and a cantilever beam problem. The performance of the CHGSO is investigated using both the manufacturing and diaphragm spring design problems taken from the automotive industry. The results obtained from using CHGSO for solving the various constrained test problems are compared with a number of established and newly invented metaheuristics, including an artificial bee colony algorithm, an ant colony algorithm, a cuckoo search algorithm, a salp swarm optimization algorithm, a grasshopper optimization algorithm, a mine blast algorithm, an ant lion optimizer, a gravitational search algorithm, a multi-verse optimizer, a Harris hawks optimization algorithm, and the original Henry gas solubility optimization algorithm. The results indicate that with selecting an appropriate chaotic map, the CHGSO is a robust optimization approach for obtaining the optimal variables in mechanical design and manufacturing optimization problems.

     

A novel hybridized metaheuristic technique in enhancing the diagnosis of cross-sectional dent damaged offshore platform members

Offshore jacket platforms are widely used for oil and gas extraction as well as transportation in shallow to moderate water depth. Tubular cross-sectional elements are used to construct offshore platforms. Tubular cross sections impart higher resistance against hydrodynamic forces and have high torsional rigidity. During operation, the members can be partially or fully damaged due to lateral impacts. The lateral impacts can be due to ship collisions or through the impact of falling objects. The impact forces can weaken some members that influence the overall performance of the platform. This demonstrates an urgent need to develop a framework that can accurately forecast dent depth as well as dent angle of the affected members. This study investigates the use of an adaptive metaheuristics algorithm to provide automatic detection of denting damage in an offshore structure. The damage information includes dent depth and the dent angle. A model is developed in combination with the percentage of the dent depth of the damaged member and is used to assess the performance of the method. It demonstrates that small changes in stiffness of individual damaged bracing members are detectable from measurements of global structural motion.     

Hybrid real-code ant colony optimisation for constrained mechanical design

This paper proposes a hybrid meta-heuristic based on integrating a local search simplex downhill (SDH) method into the search procedure of real-code ant colony optimisation (ACOR). This hybridisation leads to five hybrid algorithms where a Monte Carlo technique, a Latin hypercube sampling technique (LHS) and a translational propagation Latin hypercube design (TPLHD) algorithm are used to generate an initial population. Also, two numerical schemes for selecting an initial simplex are investigated. The original ACOR and its hybrid versions along with a variety of established meta-heuristics are implemented to solve 17 constrained test problems where a fuzzy set theory penalty function technique is used to handle design constraints. The comparative results show that the hybrid algorithms are the top performers. Using the TPLHD technique gives better results than the other sampling techniques. The hybrid optimisers are a powerful design tool for constrained mechanical design problems.     

Efficient hybrid evolutionary algorithm for optimization of a strip coiling process

This article proposes an efficient metaheuristic based on hybridization of teaching–learning-based optimization and differential evolution for optimization to improve the flatness of a strip during a strip coiling process. Differential evolution operators were integrated into the teaching–learning-based optimization with a Latin hypercube sampling technique for generation of an initial population. The objective function was introduced to reduce axial inhomogeneity of the stress distribution and the maximum compressive stress calculated by Love's elastic solution within the thin strip, which may cause an irregular surface profile of the strip during the strip coiling process. The hybrid optimizer and several well-established evolutionary algorithms (EAs) were used to solve the optimization problem. The comparative studies show that the proposed hybrid algorithm outperformed other EAs in terms of convergence rate and consistency. It was found that the proposed hybrid approach was powerful for process optimization, especially with a large-scale design problem.     

Hybrid real-code population-based incremental learning and approximate gradients for multi-objective truss design

In this article, real-code population-based incremental learning (RPBIL) is extended for multi-objective optimization. The optimizer search performance is then improved by integrating a mutation operator of evolution strategies and an approximate gradient into its computational procedure. RPBIL and its variants, along with a number of established multi-objective evolutionary algorithms, are then implemented to solve four multi-objective design problems of trusses. The design problems are posted to minimize structural mass and compliance while fulfilling stress constraints. The comparative results based on a hypervolume indicator show that the proposed hybrid RPBIL is the best performer for the large-scale truss design problems.     

An efficient optimum Latin hypercube sampling technique based on sequencing optimisation using simulated annealing

This paper proposes a new optimal Latin hypercube sampling method (OLHS) for design of a computer experiment. The new method is based on solving sequencing and continuous optimisation using simulated annealing. There are two sets of design variables used in the optimisation process: sequencing and real number variables. The special mutation operator is developed to deal with such design variables. The performance of the proposed numerical strategy is tested and compared with three established OLHS methods, namely genetic algorithm (GA), enhanced stochastic evolutionary algorithm (ESEA) and successive local enumeration (SLE). Based on 30 test problems with various design dimensions and numbers of sampling points, the proposed method gives the best results. The method can generate an optimum set of sampling points within reasonable computing time; therefore, it can be considered as a powerful tool for design of computer experiments.