Selected work

Representative Papers

The K-SCAN Clustering Algorithm

Jul 27, 2026

This work addresses the limitations of traditional clustering algorithms—specifically, the high computational complexity (O(N²)) of density-based methods like DBSCAN and the inability of partitioning approaches such as K-Means to capture nonlinear structures or handle noise effectively. To overcome these challenges, the authors propose K-SCAN, a novel algorithm that synergistically integrates vector quantization with density-based analysis. K-SCAN first employs stochastic mini-batch K-Means to generate weighted micro-clusters and then performs density connectivity analysis on these micro-clusters. This approach achieves linear time complexity while accurately identifying nonlinear manifold structures and exhibiting robustness to noise. Experimental results demonstrate that K-SCAN runs over three times faster than BIRCH on million-scale datasets, attains an Adjusted Rand Index exceeding 0.99, and remains effective even with noise levels as high as 55%.

0 citationsRead paper

Empirical Evaluation of No Free Lunch Violations in Permutation-Based Optimization

Mar 03, 2026

This study investigates under what conditions reported performance differences among optimization algorithms in benchmarking deviate from the uniform averaging assumption of the No Free Lunch (NFL) theorem within permutation-closed function spaces. By systematically applying iterative search and sampling without replacement, the work analyzes how algebraic reconstructions—via addition and subtraction—of target functions influence the ranking of algorithmic performance. Integrating correlation analysis, hierarchical clustering, Delta heatmaps, PCA, ANOVA with Tukey’s post-hoc tests, and Monte Carlo experiments, the study demonstrates that such algebraic reconstructions induce stable and statistically significant performance re-rankings, revealing localized structural biases that contradict the global symmetry implied by NFL. These reconstructed composite functions exhibit non-additive search difficulty and retain pronounced ordinal effects even in large-scale spaces, thereby supporting algorithm selection strategies grounded in problem class and objective representation.

0 citationsRead paper

A Cellular Automata Approach to Donation Game

Jul 15, 2025

This study investigates the evolutionary mechanisms of cooperation in multi-agent systems under local neighborhood interactions, focusing on the effects of environmental noise and agent strategies (e.g., reputation, generosity, tolerance). We propose a donation game modeling framework based on one-dimensional binary cellular automata, incorporating novel components: perception/action noise models, strategy mutation matrices, and agent mobility mechanisms. Our experiments demonstrate that spatial proximity significantly enhances both the emergence and long-term stability of cooperation—outperforming fully connected random interaction topologies. Moderate levels of noise improve cooperative robustness, while strategic diversity synergizes with local network structure to drive cooperative evolution. This work establishes a computationally tractable paradigm for understanding the origins of cooperation under realistic constraints—namely, limited observation capabilities and bounded interaction ranges—thereby advancing theoretical and empirical research on decentralized cooperative dynamics.

0 citationsRead paper

Recent publications

Latest Papers

The K-SCAN Clustering Algorithm

Jul 27, 2026

This work addresses the limitations of traditional clustering algorithms—specifically, the high computational complexity (O(N²)) of density-based methods like DBSCAN and the inability of partitioning approaches such as K-Means to capture nonlinear structures or handle noise effectively. To overcome these challenges, the authors propose K-SCAN, a novel algorithm that synergistically integrates vector quantization with density-based analysis. K-SCAN first employs stochastic mini-batch K-Means to generate weighted micro-clusters and then performs density connectivity analysis on these micro-clusters. This approach achieves linear time complexity while accurately identifying nonlinear manifold structures and exhibiting robustness to noise. Experimental results demonstrate that K-SCAN runs over three times faster than BIRCH on million-scale datasets, attains an Adjusted Rand Index exceeding 0.99, and remains effective even with noise levels as high as 55%.

0 citationsRead paper

Empirical Evaluation of No Free Lunch Violations in Permutation-Based Optimization

Mar 03, 2026

This study investigates under what conditions reported performance differences among optimization algorithms in benchmarking deviate from the uniform averaging assumption of the No Free Lunch (NFL) theorem within permutation-closed function spaces. By systematically applying iterative search and sampling without replacement, the work analyzes how algebraic reconstructions—via addition and subtraction—of target functions influence the ranking of algorithmic performance. Integrating correlation analysis, hierarchical clustering, Delta heatmaps, PCA, ANOVA with Tukey’s post-hoc tests, and Monte Carlo experiments, the study demonstrates that such algebraic reconstructions induce stable and statistically significant performance re-rankings, revealing localized structural biases that contradict the global symmetry implied by NFL. These reconstructed composite functions exhibit non-additive search difficulty and retain pronounced ordinal effects even in large-scale spaces, thereby supporting algorithm selection strategies grounded in problem class and objective representation.

0 citationsRead paper

A Cellular Automata Approach to Donation Game

Jul 15, 2025

This study investigates the evolutionary mechanisms of cooperation in multi-agent systems under local neighborhood interactions, focusing on the effects of environmental noise and agent strategies (e.g., reputation, generosity, tolerance). We propose a donation game modeling framework based on one-dimensional binary cellular automata, incorporating novel components: perception/action noise models, strategy mutation matrices, and agent mobility mechanisms. Our experiments demonstrate that spatial proximity significantly enhances both the emergence and long-term stability of cooperation—outperforming fully connected random interaction topologies. Moderate levels of noise improve cooperative robustness, while strategic diversity synergizes with local network structure to drive cooperative evolution. This work establishes a computationally tractable paradigm for understanding the origins of cooperation under realistic constraints—namely, limited observation capabilities and bounded interaction ranges—thereby advancing theoretical and empirical research on decentralized cooperative dynamics.

0 citationsRead paper