Mahdee Mushfique Kamal, Sujoy Das, Md. Saidur Rahman
This research addresses the challenge of generating scale-free networks within specific graph classes. It focuses on caterpillars and series-parallel networks, characterizes the value of γ for scale-free caterpillars, and presents an O(n) algorithm to generate them. It also introduces an O(m) algorithm for generating scale-free series-parallel networks.
Mahdee Mushfique Kamal, Raiyan Abdul Baten
The scientific disruption index, which tracks how publications shift focus from prior work, is a key metric in the Science of Science. Large teams, high individual productivity, and distant or international collaborations correlate negatively with disruption, while diversity, new teams, and flat structures correlate positively. This study analyzes additional pre-publication team factors to reveal how team dynamics and external influences shape the success of scientific innovation.
Mahdee Mushfique Kamal, Mohammad Al-Mahmud, Md. Saidur Rahman
This study presents the first complex network analysis of Bangladesh’s power transmission system, comprising 206 nodes and 312 weighted edges. The network shows a weak scale-free property (exponent 2.05) and low clustering (0.055), forming a tree-like topology. Centrality analysis highlights nodes with high betweenness as critical and vulnerable. Resilience testing reveals severe fragmentation risk under targeted attacks. We propose an edge-adding algorithm to improve robustness.
Hasan Masum, Mahdee Mushfique Kamal
We propose a bi-level strategy extending Autothrottle: Captain, a lightweight controller adjusting CPU via throttling signals, and Tower, a centralized bandit-based controller setting dynamic performance targets. With heuristic rules for memory scaling, our approach reduces memory limits by 50.6–60.3%, cutting waste while ensuring SLO compliance in scalable microservices.
Abu Humayed Azim Fahmid, Mahdee Mushfique Kamal
This project proposes a modified Connectome Convolutional Neural Network (CCNN) that leverages resting-state fMRI data to classify and distinguish between schizophrenia and bipolar disorder.