English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Simultaneoud model order and parameter estimation (SMOPE) was proposed for solving autoregressive exogenous system identification problem effectively using metaheuristics algorithms [1-2]. The method enabled a system’s order and parameters values to be searched simultaneously. This is possible through the way the problem is encoded in the search agents. Even though SMOPE was introduced based on particle swarm optimization (PSO) [1-2], it can easily be adapted to suit other metaheuristic algorithm such as, gravitational search algorithm (GSA) [3-4]. The PSO is a population-based optimization algorithm. The search agents of PSO, known as particles mimics how living organism such as birds and fishes look for food by exploring the search area using their own experience and information from neighborhood as guidance. The search in PSO is done iteratively. PSO’s iteration strategy can be classified as synchronous (S-PSO) and asynchronous (APSO) update [5]. S-PSO is more popular approach than APSO, where in S-PSO the movement of the whole particles in the swarm is done at once, after their performanceis evaluated. In A-PSO a particle moves as soon as its own performance is evaluated, without the need to wait for others to complete their evaluation. The direction of the movement in A-PSO is made based on whatever information available. This is a more accurate replication of nature. Random asynchronous PSO (RA-PSO) was introduced in [6]. In the original APSO the particles are evaluated and move according to the particle number. However, in RA-PSO the particle to be evaluated and move is chosen randomly, hence, in an iteration a particle can move more than once or none at all. It is found that RA-PSO is better than A-PSO. In this work the implementation of SMOPE using RA-PSO is studied and compared with SMOPE based on S-PSO. In several works, implementation of PSO with a particular iteration strategy is found to give a better result compare to other strategy. For example, Wu and Gao had reported that their adaptive inertia weight PSO implemented using asynchronous update has a better performance than the same approach implemented using synchronous update [7]. In [8], APSO with discrete crossover is found to perform better than S-PSO with the crossover operator. However, Engelbrecht in his work concluded that there is no definite winner of S-PSO vs A-PSO but rather it is a function dependent option [5]. The same observation is made in [9]. Therefore, in this work the performance of RAPSO based SMOPE is compared with the S-PSO based SMOPE. Six ARX system identification problems are used. The results show that RA-PSO on average has a slightly better performance.

Simultaneous Model Order and Parameter Estimation (SMOPE) based on Random Asynchronous Particle Swarm Optimization