Benchmark Results for Sparse Matrix Optimization Across 237 Test Problems

Abstract

sparse matrix optimization is a central problem in applied mathematics, yet standard techniques scale poorly via citizen-science data. We assemble a dataset of 4631 observations and evaluate three competing models under matched conditions. Our measurements agree with theoretical predictions to within 15%, resolving an open discrepancy in the literature. We discuss limitations and outline directions for follow-up work at the undergraduate level.