Browsing by Author "Fadlelseed, Sajid"

Now showing 1 - 2 of 2

Artificial Ecosystem‑Based Optimization with Dwarf Mongoose Optimization for Feature Selection and Global Optimization Problems
(International Journal of Computational Intelligence Systems, 2023) Al‑Shourbaji, Ibrahim; Kachare, Pramod; Fadlelseed, Sajid; Jabbari, Abdoh; Hussien, Abdelazim G.; Al‑Saqqar, Faisal; Abualigah, Laith; Alameen, Abdalla
Meta-Heuristic (MH) algorithms have recently proven successful in a broad range of applications because of their strong capabilities in picking the optimal features and removing redundant and irrelevant features. Artificial Ecosystem-based Opti mization (AEO) shows extraordinary ability in the exploration stage and poor exploitation because of its stochastic nature. Dwarf Mongoose Optimization Algorithm (DMOA) is a recent MH algorithm showing a high exploitation capability. This paper proposes AEO-DMOA Feature Selection (FS) by integrating AEO and DMOA to develop an efficient FS algorithm with a better equilibrium between exploration and exploitation. The performance of the AEO-DMOA is investigated on seven datasets from different domains and a collection of twenty-eight global optimization functions, eighteen CEC2017, and ten CEC2019 benchmark functions. Comparative study and statistical analysis demonstrate that AEO-DMOA gives competi tive results and is statistically significant compared to other popular MH approaches. The benchmark function results also indicate enhanced performance in high-dimensional search space.
LBP-CA: A Short-term Scheduler with Criticality Arithmetic
(University of Hertfordshire, 2022) Fadlelseed, Sajid; Kirner, Raimund; Menon, Catherine
In safety-critical systems a smooth degradation of services is a way to deal with resource shortages. Criticality arithmetic is a technique to implement services of higher criticality by several tasks of lower criticality. In this paper, we present LBP-CA, a mixed-criticality scheduling protocol with smooth degradation based on criticality arithmetic. In the experiments we show that LPB-CA can schedule more tasks than related scheduling protocols (BP and LBP) in case of resource shortage, minimising the negative effect on low-criticality services. This is achieved by considering information about criticality arithmetic of services.