The approach is based on a meta-heuristic optimisation technique to perform the optimization. Commonly used algorithms are genetic algorithms and particle swarm optimization are often incapable of reaching global minimum and the computational time and effort makes them impractical. Nelder-Mead optimization is used in a two-stage process taking inexpensively obtained approximate solutions from these algorithms and refining it to obtain near global minimum of a given error function. The initial solutions (vertices) are required for the application of Nelder-Mead optimization. This hybrid method is better than individual algorithms, making the approach ideal for integration into the six-sigma process in manufacturing where ease of deployment and rapid results are essential.