TY - CHAP

T1 - Soft Optimal Computing to Identify Surface Roughness in Manufacturing Using a Gaussian and a Trigonometric Regressor

AU - Haus, Benedikt

AU - Mercorelli, Paolo

AU - Yap, Jin Siang

AU - Schäfer, Lennart

PY - 2022

Y1 - 2022

N2 - This contribution deals with the identification of roughness as a function of gloss in manufacturing using Particle Swarm Optimization (PSO) methods. The proposed PSO method uses a Least Squares Method as a cost function to be optimized. The identification structure uses a Gaussian and a Trigonometric Regressor characterized by seven parameters to be estimated. In PSO algorithms, there is a delicate balance to maintain between exploration (global search) and exploitation (local search) and this is one of the most important issues of this optimization method. An analysis of an increment of the dimension of the search space of the PSO is proposed. This is realized through an increment of its exploitation dimension to improve the precision of the search phase of the PSO, at the cost of more computations in each iteration. Nevertheless, convergence time results to be shorter in the presented case. Thus, an optimal increment of the dimension exists which states a compromise between velocity of the convergence and precision. Measured results from a manufacturing system with and without enlargement of the search space are shown together with results obtained using a Genetic Algorithm (GA) for comparison. Advantages and drawbacks are pointed out.

AB - This contribution deals with the identification of roughness as a function of gloss in manufacturing using Particle Swarm Optimization (PSO) methods. The proposed PSO method uses a Least Squares Method as a cost function to be optimized. The identification structure uses a Gaussian and a Trigonometric Regressor characterized by seven parameters to be estimated. In PSO algorithms, there is a delicate balance to maintain between exploration (global search) and exploitation (local search) and this is one of the most important issues of this optimization method. An analysis of an increment of the dimension of the search space of the PSO is proposed. This is realized through an increment of its exploitation dimension to improve the precision of the search phase of the PSO, at the cost of more computations in each iteration. Nevertheless, convergence time results to be shorter in the presented case. Thus, an optimal increment of the dimension exists which states a compromise between velocity of the convergence and precision. Measured results from a manufacturing system with and without enlargement of the search space are shown together with results obtained using a Genetic Algorithm (GA) for comparison. Advantages and drawbacks are pointed out.

KW - Curve fitting

KW - Manufacturing applications

KW - Particle Swarm Optimization

KW - Engineering

UR - http://www.scopus.com/inward/record.url?scp=85130279372&partnerID=8YFLogxK

UR - https://www.mendeley.com/catalogue/43901d95-2f2d-3bd1-a0f1-8e1d540724ff/

U2 - 10.1007/978-3-031-05767-0_4

DO - 10.1007/978-3-031-05767-0_4

M3 - Article in conference proceedings

AN - SCOPUS:85130279372

SN - 978-3-031-05766-3

T3 - Communications in Computer and Information Science

SP - 41

EP - 50

BT - Soft Computing and its Engineering Applications

A2 - Patel, Kanubhai K.

A2 - Doctor, Gayatri

A2 - Patel, Atul

A2 - Lingras, Pawan

PB - Springer Nature Switzerland AG

CY - Cham

T2 - 3rd International Conference on Soft Computing and its Engineering Applications - icSoftComp 2021

Y2 - 10 December 2021 through 11 December 2021

ER -