A multi-population particle swarm optimization algorithm with adaptive patterns of movement for the stochastic reconstruction of petroleum fractions

•A novel optimization algorithm (MPSO-AMP) is proposed by combining multipopulation heuristic and movement pattern adaptation heuristic.•In terms of the optimum ratio, which evaluates the best solution, the proposed algorithm achieved results ranging from 0.30 to 1.02.•An assessment is also presente...

Full description

Saved in:
Bibliographic Details
Published in:Computers & chemical engineering 2023-06, Vol.174, p.108221, Article 108221
Main Authors: Dantas, T.S.S., Noriler, D., Huziwara, K.W.
Format: Article
Language:English
Subjects:
Compositional modeling
Molecular reconstruction
Particle swarm optimization
Petroleum feedstock
Stochastic reconstruction
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•A novel optimization algorithm (MPSO-AMP) is proposed by combining multipopulation heuristic and movement pattern adaptation heuristic.•In terms of the optimum ratio, which evaluates the best solution, the proposed algorithm achieved results ranging from 0.30 to 1.02.•An assessment is also presented on the performance in Stochastic Reconstruction of several metaheuristic extensions of the PSO algorithm.•The results indicate that the use of metaheuristics in solving the molecular reconstruction problem can be advantageous, depending on the molecular model and experimental data. Molecular reconstruction (MR) methods are used to estimate the detailed composition of a hydrocarbon mixture at the molecular level based on experimental data. Stochastic reconstruction (SR) is one of the MR approaches available to this end. Obtaining the solution involves estimating a set of probability density function (PDF) parameters with a global optimization algorithm. However, to the author's knowledge, no other study has thoroughly explored using metaheuristics to improve parameter estimation in this problem. By combining ideas from Chang (2015) and Bonyadi (2018), this study proposes a novel heuristic optimization algorithm named Multi-Population Particle Swarm Optimization with Adaptive Movement Patterns (MPSO-AMP). Since multiple swarms are employed simultaneously with unique movement patterns, the proposed algorithm can render more exact solutions in SR problems than conventional particle swarm optimization (PSO) and other extensions from the literature.
ISSN:0098-1354
1873-4375
DOI:10.1016/j.compchemeng.2023.108221