Design and Development of an Improved Flower Pollination Algorithm

Koay, Brenda Chia Wen (2022) Design and Development of an Improved Flower Pollination Algorithm. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Brenda Koay Chia Wen.pdf Restricted to Registered users only Download (2MB)

Abstract

Since a long time ago, the process of solving problem has been inspired by nature and many biological systems had evolved especially in the aspect of reproduction process efficiency over the time. Hence, many algorithms were also created from the inspiration of effective characteristic of nature’s biological process. Among all the nature-inspired algorithms, a novel algorithm namely the flower pollination algorithm (FPA) introduced by Yang in the year of 2012 had shown results with relatively higher frequency compared to other techniques especially improving the classification results of a system after being implemented as a classifier of that system. In the other hand, Malaysia as the main contributor of the world’s palm oil production need to cope with the high demand for oil palm exports where the ripeness level of the oil palm fruit should be taken into account so that optimum palm oil could be extracted. The current implementation of manual classification brings drawbacks such as human vision observation process is time-consuming, classification results are inconsistent and may be inaccurate due to different perspective from various personnel. These problems can be solved by incorporating an intelligence system which can classify the oil palm fruitlets according to the extracted datasets from oil palm fruit images automatically. Thus, in this paper studies on implementation of FPA into classification of oil palm fruitlets in Malaysia will be done and the improved version of FPA classifier by incorporating SCA together will be designed to further improve the performance of the oil palm fruitlets ripeness classification.