Integration of technique for order preference by similarity to ideal solution and back-propagation neural network/genetic algorithm to optimise light guide plate surface roughness process parameters

Abstract

With the rapid development of liquid crystal display technology, product design focuses on high quality, high resolution, thin and lightweight specifications and uniform display brightness. However, the liquid crystal display thickness is determined by the design of the backlight module light source layout. Different light source designs with different optical materials can realise different optical effects. Hence, this study discusses the impact of the surface roughness of backlight module light guide plate on light transmission. With an edge backlight module as an example, this study assumes that the incident surface roughness of the light guide plate depends on control of CNC processing parameters, using the technique for order preference by similarity to ideal solution and back-propagation neural network/genetic algorithm to identify key factors from relevant factors, followed by validation of the optimisation parameter process capability. The roughness Cpk process capability is improved from the original 0.5 to 1.75.