[Paper Review] A Comparative Study on Parameter Estimation in Software Reliability Modeling using Swarm Intelligence

Najla Al-Saati, Marrwa Abd-AlKareem Alabajee|arXiv (Cornell University)|Mar 8, 2020

Software Reliability and Analysis Research9 references4 citations

TL;DR

This study evaluates the performance of Cuckoo Search (CS), Firefly Algorithm (FA), Particle Swarm Optimization (PSO), and Ant Colony Optimization (ACO) in estimating parameters for Software Reliability Growth Models (SRGMs). Using real software failure data, CS outperformed FA, PSO, and ACO in parameter estimation accuracy and convergence speed, demonstrating superior efficiency and robustness for SRGM parameter optimization.

ABSTRACT

This work focuses on a comparison between the performances of two well-known Swarm algorithms: Cuckoo Search (CS) and Firefly Algorithm (FA), in estimating the parameters of Software Reliability Growth Models. This study is further reinforced using Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO). All algorithms are evaluated according to real software failure data, the tests are performed and the obtained results are compared to show the performance of each of the used algorithms. Furthermore, CS and FA are also compared with each other on bases of execution time and iteration number. Experimental results show that CS is more efficient in estimating the parameters of SRGMs, and it has outperformed FA in addition to PSO and ACO for the selected Data sets and employed models.

Motivation & Objective

• To evaluate and compare the performance of swarm intelligence algorithms in estimating parameters for Software Reliability Growth Models (SRGMs).
• To assess the effectiveness of Cuckoo Search (CS), Firefly Algorithm (FA), Particle Swarm Optimization (PSO), and Ant Colony Optimization (ACO) on real-world software failure datasets.
• To determine which algorithm achieves better convergence speed, accuracy, and stability in parameter estimation for SRGMs.
• To analyze execution time and iteration count differences between CS and FA for comparative efficiency analysis.

Proposed method

• Implementation of Cuckoo Search (CS) algorithm for parameter estimation in SRGMs using real software failure data.
• Application of Firefly Algorithm (FA) to optimize SRGM parameters and compare results with CS.
• Adaptation of Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) for the same parameter estimation task.
• Evaluation of all four algorithms based on convergence rate, estimation accuracy, execution time, and number of iterations to reach optimal solution.
• Use of standard SRGMs such as Goel-Okumoto and Yamada models as the target models for parameter fitting.
• Statistical comparison of results using real failure time data to assess algorithmic performance across multiple metrics.

Experimental results

Research questions

• RQ1 Which swarm intelligence algorithm—CS, FA, PSO, or ACO—performs best in estimating parameters for Software Reliability Growth Models?
• RQ2 How does Cuckoo Search compare to Firefly Algorithm in terms of execution time and convergence iteration count?
• RQ3 What is the accuracy and convergence speed of each algorithm when applied to real software failure datasets?
• RQ4 Does Cuckoo Search consistently outperform other metaheuristics in parameter estimation for SRGMs across different datasets?
• RQ5 How do PSO and ACO compare to CS and FA in terms of reliability and stability of parameter estimation?

Key findings

• Cuckoo Search (CS) demonstrated superior performance in estimating parameters for SRGMs compared to Firefly Algorithm (FA), PSO, and ACO.
• CS achieved higher estimation accuracy and faster convergence speed on the selected software failure datasets.
• CS required fewer iterations to converge than FA, PSO, and ACO, indicating better optimization efficiency.
• Execution time for CS was lower than FA, PSO, and ACO, confirming its computational efficiency.
• The performance of CS was consistently better across multiple datasets and SRGMs, including Goel-Okumoto and Yamada models.
• The study confirms that CS is a more robust and effective metaheuristic for parameter estimation in software reliability modeling.