Sistem Pakar Pemupukan Kelapa Sawit Menggunakan Metode Forward Chaining

Abstract

Oil palm smallholders require knowledge of fertilization because the fertilization is an influential factor in the productivity of fresh fruit bunches. On the other hand, the oil palm smallholders generally don’t have fertilizer expertise and they face a difficulty in determining good fertilization (good in type, dosage, time and manner) based on various literatures. Therefore, a simple application that can provide recommendations for palm oil fertilization is developed. The recommendations are based on expert knowledge that have been written in various literatures. Like an expert system, the application is designed by compiling knowledge base, constructing an inference engine and creating a prototype. The knowledge bases, that are facts and rules, are prepared using rule-based reasoning approach, the inference engine is compiled using forward chaining methods and the prototype is made using a visual basic application in a spreadsheet application. The application design uses facts of plant varieties, phases or age of plants, soil types, nutrient deficiency in soil and plants, rainfall, fertilizer content, and nutrient properties (synergies or antagonist) on fertilizers. The Applications made capable of providing fertilization recommendations by conducting forward chaining based on data-given facts and rules that have been made on the knowledge base.