Software optimization for input converter utility tool software using deep learning technique

Abstract

The world continue to move very fast when it comes to technology, different field of industry continues to develop and apply many new found ideas and concept that will help ease and gives best throughput results in their corresponding production area. One sample firm of this industry is “Finisar” in which the contribution of this research work will give credit. This thesis will study the problem how to optimize and what optimization method that can suit to improve and mitigate the errors encountering to one of Finisar software tool which is called Horsham Converter Tool. It will also study what is the differences and improvement between the current and the proposed optimized software. Upon completing this research, it should be identified already what optimization methods that will be used to lower the risk in error and enhance the current process for converting Specification File. This work also aiming to evaluate and show the comparison between the current and propose optimized method using the actual software simulation. The method proposed is the deep learning technique which is one field in AI algorithm and will also discuss other different related fields in AI that can contribute clarification on this research. Also it add some data structure concept in processing the array data which additional optimization of the software tool. The result of the simulation of the research was measured by accuracy and the time lapse or the execution time. From accuracy of 29.40% in old method to 94.44 % to the new proposed method was achieved upon simulation of the proposed method. Upon simulation of the new Input Converter Utility Tool Software, different product codes was tested and gives the different accuracy and its time lapse or the execution time. It shows from the tabulated table that the accuracy varies accordingly and directly proportional to its amount of data that was used.