Decision support solution for water quality management in shrimp aquaculture

Abstract

Aquaculture industry is one ofthe booming industries in today’s world. Since there is an appropriate environmental condition high amount of natural resources availability, positive socio economic impact and due to the great potentials for the development of aquaculture it popular in Sri Lanka also. Still expected target is far beyond due to various reasons such as failure of the harvest, high mortality, less growth and uncertainty ofthe production. Researchers have found that one ofthe major reason for the above mentioned problems are lack ofmanagement practices in the industry. This project focuses on water quality management one of the key area in intensive shrimp farming. As an initial step this project is dedicated to introduce automated tool for proper data collection and timely accurate decision support for non-expert users. Permanent, stable data storage to store data for future decision making process is another advantage ofthe project. Given solution contains two main sections. Smart phone application and decision making module. Around 2000 past records containing water quality parameters, observations and decisions and recommendations given by expert is analyzed to identify any past pattern. K-means clustering mechanism is used to group similar cases together and merged those groups with relevant decision and recommendation. When the new case comes system uses past experience to identify the new situation and help quick decision making process. Field workers input water quality parameters and observations using mobile interface. Collected data from different ponds transfer to central database through web server. Trained system process data to produce current pond status and recommendations as an output. This will help non expert users to get immediate attention over ponds. Use cross validation for the evaluation ofan algorithm .system testing is done using 500 records of current culture to test the system. Reliable fast remote data collection and decision support system for non-expert users have been implemented and at the same time implementation contributes to bridge the Information technology gap in the field ofshrimp farming