Simulation Verification Comparing Pseudo-IGM and Non-Linear WLS Optimization
Source:R/simulations.R
runWLSEOpt.RdExecutes a Monte Carlo simulation benchmarking the closed-form priority weight vectors
derived by Pseudo-IGM (pseudoIGM) against the numerical global optimization
solver (WLSopt) via nloptr, verifying that closed-form IGM formulas
consistently achieve the exact global minimum Weighted Least Squares Error (WLSE) without
relying on iterative numerical solvers (Section 4.2; Yuen, 2024).
Arguments
- size
Integer. Total number of random PRM simulation instances to evaluate (default:
10).- seedNo
Integer. Seed number for pseudo-random number generator initialization to ensure reproducibility (default:
100).- sigDigits
Integer. Number of significant decimal places used for rounding priority weight vectors during equality verification (default:
4).- errorFile
Character string. Output text file path for logging simulation results and discrepancy metrics (default:
"wlseOptData.txt").
Value
Logical. Returns FALSE if all simulation iterations pass without discrepancy, or TRUE if a discrepancy is detected.
References
Yuen, K. K. F. (2024). Inverse Gram Matrix Methods for Prioritization in Analytic Hierarchy Process: Explainability of Weighted Least Squares Optimization Method. arXiv preprint arXiv:2401.01190. https://doi.org/10.48550/arXiv.2401.01190
Examples
if (FALSE) { # \dontrun{
# Run simulation comparing Pseudo-IGM against numerical solver for 5 random PRMs
runWLSEOpt(size = 5, seedNo = 123, errorFile = tempfile(fileext = ".txt"))
} # }