|
Tomato fruit is one of agroproducts that has high-economic value in the world particularly inIndonesia. To compete in a worldwide market a tomato fruit producer must produce fresh or processedtomato with high quality. High quality tomato products are influenced by the application of post-harvesttreatment or processing. One of the vital process in post-harvest treatment is sortation. Mannualsortation introduces subjectivity (bias), inaccuracy, slowness and inconsistency. This needs moreintelligent sortation methods and tools that overcome the sort comings of manual process. ProbabilisticNeural Network (PNN) is one of Artificial Neural Network (ANM variants that can be to develop acomputer-based sortation engine for tomato fruits. However, to accelerate the sortation process, parallelcomputation is employed allowing multiple processors to execute simultaneously the sortation process.This research is aimed towards the implementation and testing of a parallel computation algorithm withPNN to perform sortation for tomato fruits. Some criteria being observed and tested include accuracy,total execution time, speedup, and efficiency compared to sequential algorithm. The experimental resultsshow that the application of parallel computation algorithm with PNN introduces the increase ofaccuracy, total execution time, speedup, and efficiency with the same accuracy. |
