Convergence Analysis of Swarm Intelligence Metaheuristic Methods

Davidović, Tatjana; Krüger, Tatjana Jakšić

doi:10.1007/978-3-319-93800-4_20

Cited by 2 publications

(2 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While t < T Calculate the fitness, find the best solution and update ES using Equation (6). For i = 1 : N For j = 1 : n Update η (i,j) , R (i,j) , P (i,j) , and M(x i ) using Equations ( 4), ( 5), ( 7) and (8).…”

Section: Algorithm 1: Pseudocode Of the Reptile Search Algorithmmentioning

confidence: 99%

“…These factors pose significant challenges in addressing the feasibility and economic considerations of actual situations. In contrast, heuristic algorithms encompass several approaches, such as greedy strategies and local search algorithms [7][8][9]. These algorithms rely on the inherent laws of the problem to obtain improved workable solutions.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

MLBRSA: Multi-Learning-Based Reptile Search Algorithm for Global Optimization and Software Requirement Prioritization Problems

Kailasam,

Nalliah,

Nallagoundanpalayam Muthusamy

et al. 2023

Biomimetics

View full text Add to dashboard Cite

In the realm of computational problem-solving, the search for efficient algorithms tailored for real-world engineering challenges and software requirement prioritization is relentless. This paper introduces the Multi-Learning-Based Reptile Search Algorithm (MLBRSA), a novel approach that synergistically integrates Q-learning, competitive learning, and adaptive learning techniques. The essence of multi-learning lies in harnessing the strengths of these individual learning paradigms to foster a more robust and versatile search mechanism. Q-learning brings the advantage of reinforcement learning, enabling the algorithm to make informed decisions based on past experiences. On the other hand, competitive learning introduces an element of competition, ensuring that the best solutions are continually evolving and adapting. Lastly, adaptive learning ensures the algorithm remains flexible, adjusting the traditional Reptile Search Algorithm (RSA) parameters. The application of the MLBRSA to numerical benchmarks and a few real-world engineering problems demonstrates its ability to find optimal solutions in complex problem spaces. Furthermore, when applied to the complicated task of software requirement prioritization, MLBRSA showcases its capability to rank requirements effectively, ensuring that critical software functionalities are addressed promptly. Based on the results obtained, the MLBRSA stands as evidence of the potential of multi-learning, offering a promising solution to engineering and software-centric challenges. Its adaptability, competitiveness, and experience-driven approach make it a valuable tool for researchers and practitioners.

show abstract

Section: Algorithm 1: Pseudocode Of the Reptile Search Algorithmmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%