Application of fuzzy logic in decision-making process for relocation of floating net cages in river fish farming

Land-based aquaculture operations, at present,  are intensively conducted to meet the ever-growing demand for food consumption. Floating net cages are one of the traditional methods commonly used by Indonesian fishermen for river fish farming. Increased human activities along the Musi River and coastline have resulted in pollution and waste in the river waters and fluctuating water quality. Yet, floating net cage owners still manually assess the water quality. This study aims to develop an early warning system for water quality and create a decision-making program as a reference for fishermen to relocate floating net cages when the river water quality deteriorates.

The device was tested at 39 locations within a radius of approximately 3400 meters, and the distance between locations varied between 55 and 334 meters. The river was divided into three sections: the river coast, the middle section, and the other river coast. Water quality sensors were placed at a depth of 0–20 centimeters from the surface of the Musi River, with measurement durations at each location ranging from 1 to 40 minutes. Direct measurements of the Musi River's water quality were obtained by monitoring the water quality using an internet-based computer application. A decision-making Python program utilizing fuzzy logic was then executed to evaluate the suitability of the river water quality for fish cultivation. The program's input variables comprise water temperature, potential of hydrogen, and dissolved oxygen sensor data. Meanwhile, the program output recommends floating net cage owners to either "Stay in position" or "Move." Water quality warnings that exceed the upper and lower threshold limits are displayed using light-emitting diode indicators and a buzzer.

Overall, the water quality values of the Musi River at the test locations generally indicated stable and suitable conditions for river fish cultivation. The average water quality values were 29.20 degrees Celsius for temperature, 3.98 milligrams per liter for dissolved oxygen, and a potential of hydrogen of 6.42. From all the data obtained during the decision-making program, 36 locations suggested that the floating net cages should "Stay in position." Meanwhile, the three remaining locations were recommended to "Move" as they exhibited poor water quality, with potential of hydrogen values below 6. Field observations indicated that these locations were situated near residential areas, factories/industries, and tributaries, which are highly susceptible to waste and pollution. The output of the decision-making program correlated with the issued warnings by the water quality warning indicators when the pH value exceeded the lower threshold limit.

The fuzzy logic method implemented in the Python program for decision-making regarding the relocation of floating net cages in river fish farming revealed the fluctuating water quality conditions of the Musi River within a specific time duration. These conditions correlated with the proximity of the water bodies to pollution sources such as residential areas, factories, and tributaries. The program's output classified the status of the floating net cages into two conditions: "Stay in position" or "Move." The decision-making application to relocate floating net cages for fish farming in rivers provides a solution for fishermen as the resulting program decisions give the same indication as the reading value of the water quality sensor.