Adaptive Anchor-Guided Spherical-Neutrosophic Intelligence for Uncertainty-Aware Engineering Systems — Research Software and Reproducibility Package

Abstract

This repository provides the research software, reproducibility materials, benchmark results, and figure-generation resources accompanying the study “Adaptive Anchor-Guided Spherical-Neutrosophic Intelligence for Uncertainty-Aware Engineering Systems.” The package implements and supports the proposed Adaptive Anchor-Guided Spherical-Neutrosophic (A-AGSN) framework for uncertainty-aware intelligent engineering systems. It includes Python source code, benchmark and experimental resources, structured results, figure-generation scripts, and documentation required to reproduce and inspect the principal computational components reported in the associated manuscript. The released materials cover anchor-guided fuzzy clustering (A-FCM), spherical-neutrosophic uncertainty representation, reliability and risk-reduction analysis, the tabular reinforcement-learning proof-of-concept, Dynamic Anchor Protocol experiments, Dynamic Anchor-Based Quantization (DAQ), hyperparameter sensitivity analysis, and the generation of manuscript figures. The clustering component is evaluated on six benchmark datasets: UCI Iris, Wine, Breast Cancer Wisconsin, Synthetic Multi-Sensor, MotionSense-HAR, and Occupancy Detection. The reinforcement-learning component is provided as a controlled low-dimensional tabular proof-of-concept, while the Dynamic Anchor Protocol is evaluated through a standalone synthetic waypoint experiment. DAQ hardware speedups reported in the associated study are analytical cycle-count-based projections; the package provides software-emulation results rather than physical ARM/FPGA hardware measurements. The archive also contains documentation and metadata intended to facilitate transparent reuse, verification, citation, and extension of the proposed framework. Version: 1.0.0