Željko Kanović

Time-varying PSO -- convergence analysis, convergence-related parameterization and new parameter adjustment schemes

In this paper, a formal convergence analysis of the conventional PSO algorithms with time-varying parameters is presented. Based on this analysis, a new convergence-related parametric model for the conventional PSO is introduced. Finally, several new schemes for parameter adjustment, providing significant performance benefits, are introduced. Performance of these schemes is empirically compared to conventional PSO algorithms on a set of selected benchmarks. The tests prove effectiveness of the newly introduced schemes, especially regarding their ability to efficiently explore the search space.

Generalized particle swarm optimization algorithm - Theoretical and empirical analysis with application in fault detection

A generalization of the particle swarm optimization (PSO) algorithm is presented in this paper. The novel optimizer, the Generalized PSO (GPSO), is inspired by linear control theory. It enables direct control over the key aspects of particle dynamics during the optimization process. A detailed theoretical and empirical analysis is presented, and parameter-tuning schemes are proposed. GPSO is compared to the classical PSO and genetic algorithm (GA) on a set of benchmark problems. The results clearly demonstrate the effectiveness of the proposed algorithm. Finally, an application of the GPSO algorithm to the fine-tuning of the support vector machines classifier for electrical machines fault detection is presented.

HCR Gearing and Optimization of its Geometry

A special kind of the basic involute profile of non-standard gearing is called high contact ratio (HCR) gearing, where the contact ratio is higher, there are always at least two pairs of teeth in contact and the unit addendum height is not equal to one like for standard gearing. Thus, the tooth height is increased. When HCR gearing is used, it is not necessary to achieve a greater gear load capacity, but nevertheless there is a greater risk of interference due to the greater tooth height. The advantages of HCR gearing is higher resistance (load is distributed on more pairs of teeth at the same time) and a lower relative noise level of gearing, which can be significantly reduced by using an integer HCR factor. HCR profiles are more complicated than standard involute profiles, they have a greater predisposition for interference, pointed tip thickness and undercut of teeth during production (primary production interference). Due to increased addendum height, there is a larger possibility of some interference or pointed tooth tip occurring. Therefore, these issues need to be prevented in the design phase, and ensured that all relevant equations and constraints are satisfied. The described method of finding optimal gear parameter values uses a Generalized Particle Swarm (GPS) optimization algorithm and MATLAB. The GPS optimization is shown to be a very fast and reliable method.

HCR Gearing Geometry Optimization by Using of Generalized Particle Swarm Algorithm

Pitting and temperature scuffing evidenced by damage to teeth flanks of gears are the most important problems needing to be solved in the process of gearing design and calculation. According to current valid standards, such calculations can be resolved with a high level of reliability for all the usual gearing types. However, suitable calculations for High Contact Ratio (HCR) gears have not been adequately researched to date. It has been identified that in HCR gears some different process of scuffing and pitting formation occurs during the gear’s operation. Therefore it is actual to deal with optimization of HCR gearing just from these kinds of teeth flanks damages. In article, the authors describe a new method for finding optimal solutions for geometrical parameters \( h_{a1}^{*} ,h_{a2}^{*} \) and x1, using a Generalized Particle Swarm Optimization Algorithm for both above mentioned optimization tasks.

Application of Nature-Inspired Optimization Techniques in Vessel Traffic Control

This chapter aims to present the analysis and comparison of some well-known nature-inspired global optimization techniques applied to an expert system controlling a ship locking process. A ship lock zone represents a specific area on waterway, and control of the ship lockage process requires a comprehensive approach. The initially proposed Fuzzy Expert System (FES) was developed using suggestions obtained from lockmasters (ship lock operators) with extensive experience. Further optimization of the membership function parameters of the input variables was performed to achieve better results in the local distribution of vessel arrivals. The purpose of the analysis and comparison is to find the best algorithm for optimization of membership functions parameters of FES for the ship lock control. The initially proposed FES is optimized (fine-tuned) with three global optimization algorithms from the group of evolutionary and swarm intelligence algorithms, in order to achieve the best value of the economic criterion defined as a linear combination of two opposed criteria: minimal average waiting time per vessel and minimal number of empty lockages (lockages without a vessel in a chamber). Besides the well known and widely applied Genetic Algorithm (GA), two relatively new but very promising global optimization techniques were used: Particle Swarm Optimization (PSO), the technique based on behavior of animals living in swarms and Artificial Bee Colony (ABC) algorithm, inspired by social organization of honey bees. Although all these algorithms have been widely applied and showed a great potential in engineering applications in general, their application in ship lock control and similar transportation problems is not so common. However, this chapter will present that all three algorithms may obtain the significant improvement of the adopted economic criterion value and succeed to find its (possibly global) optimum. Furthermore, the performances of these algorithms in FES parameters optimization are compared and some conclusions are adopted on their applicability, efficiency, and effectiveness in similar systems. The developed fuzzy algorithm is a rare application of artificial intelligence in navigable canals and significantly improves the performance of the ship lockage process. This adaptable FES is designed to be used as a support in decision-making processes or for the direct control of ship lock operations.

Optimization of HCR Gearing Geometry from a Scuffing Point of View

The paper deals with an issue of increasing the resistance of HCR external involute gearing from a scuffing point of view. It reports on a difference between involute gears with low contact ratio (LCR) and those with high contact ratio (HCR). The paper describes scuffing as the most significant damage done to the teeth flanks of HCR involute gears. In the case of warm scuffing, it is the combined action of high pressure between surfaces, high sliding speeds, and excessive contact temperature, resulting from pressure and sliding speed values, which consequently cause oil film rupture between the teeth flanks. Adopting a suitable geometry of the tooth curve profile, certain values of addendum heights for the meshing wheel will be defined according to the criteria of specific slips and corrected head shapes of the teeth of both wheels. The paper deals with the assessment and theoretical analysis of the impact of the HCR tooth profile resistance to scuffing on the basis of integral temperature criterion according to the Winter-Michaelis criterion. The basic relations for assessing the scuffing properties of an HCR involute gearing profile are derived, as well as the optimization of geometrical parameters of HCR gearing based on theoretical considerations on the properties of HCR gearing in terms of its resistance to warm scuffing combined with geometrical constraints against interferences. Finally, the results of the obtained optimization are compared with experimental results provided in previous research. A significant benefit in a theoretical area is the generalization of the integral temperature criterion for involute HCR gearing. By optimizing the criterion for the integral temperature of involute HCR gearing, the minimal flash temperature is obtained and the condition for the occurrence of scuffing is minimized.