Hybrid Algorithm with Weighted Nonlinear Glial Ratio Neural Networks for Coal Mine Rescue Operation

In this paper a Glial Ratio (g-ratio) mix hybrids of 67% Sigmoid and 33% Radial functions (HSCR-BFgr) based on Particle swarm optimisation with the highest survivability of all possible routing redundancies, reliability, efficiency, fault tolerant with minimum fitness error is proposed for underground rescue operation. Nonlinear weights of cosine and sine were imposed on the g-ratio hybrids. In addition we introduced a nonlinear weight with the g-ratio on the Gaussian RBF. The performance of the Hybrid with negative cosine weight (HSCR-BFgrcos) was the best among the various g-ratio hybrids as compared to Gaussian with the same nonlinear weight. The hybrid with negative nonlinear cosine weight yielded the best results with an optimised error of 0.011. The proposed Nonlinear Hybrid Algorithm has better capability of approximation to underlying functions with a fast learning speed, high scalability, robusticity and is competitive to the Gaussian with the same nonlinear weight.