Stanislus29/stangorithms: Geometric Clustering Minimization with Recursive QM Optimization

Release 2.1.0: Geometric Clustering Minimization with Recursive QM Optimization This release introduces geometric clustering algorithms that enhance hierarchical Boolean minimization by exploiting spatial relationships within K-map representations. We partition the minimization problem into geometrically coherent subspaces, dramatically reducing the search space for subsequent Quine-McCluskey (QM) optimization. Core Methodology The geometric preprocessing stage identifies and merges spatially adjacent minterms through dimensional projection, producing an optimized set of prime implicants for the recursive QM engine. This hybrid approach combines the visual structure of K-maps with the algorithmic rigor of tabular methods, achieving substantial computational advantages over purely algebraic techniques. Theoretical Findings Extensive empirical analysis reveals distinct decay characteristics and fundamental complexity limits in geometric clustering: 3D Functions: Decay occurs at n > 8 variables 4D Functions: Decay occurs at n > 10 variables Our forthcoming paper demonstrates that geometric approaches exhibit structural collapse beyond 10 variables, establishing theoretical bounds on the applicability of spatial clustering methods for high-dimensional Boolean minimization. Performance Benchmarks Comparative analysis against PyEDA's espresso_exprs() reveals exceptional efficiency: 3D Functions: 4% deviation at maximum variable count (n = 8) 4D Functions: 6% deviation at maximum variable count (n = 10) These metrics demonstrate near-optimal minimization quality while maintaining computational tractability for practical circuit design. Broader Implications This release bridges geometric intuition with algebraic optimization, achieving state-of-the-art performance in logic synthesis. It provides foundations for understanding information loss in Boolean autoencoders and the topological implications of information structures.