An expert system for planning and designing dairy farms in hot climates

Abstract

Eleven simulation models were developed to plan and design several dairy farm facilities.  Subsequently, an electronic spark map (decision tree) was developed for each simulation model, and then the simulation models were integrated into the relevant spark maps.  Afterwards, C# language (C Sharp), which is an object-oriented programming language, was used to develop an expert system via the simulation models and the electronic spark maps.  The developed expert system is able to plan and design several dairy farm facilities, e.g. housing system (corrals system), shade structure and roof material, concrete base, cooling system, milking parlour, forage storage, and manure handling system.  Subsequently, it plans the farmstead layout, and it leads to implement the technologies, equipments, and machines required for performing several farm operations.  Furthermore, it studies water and electricity requirements of the planned dairy farm and the available sources on site.  Moreover, it calculates the capital investment and the fixed, variable, and total costs.  Data of six dairy farms were used to carry out the expert system validation and evaluation.  The differences between the actual and calculated values were determined and the standard deviations were calculated.  The coefficients of variation range between 3% and 7%.  The accuracy of the developed expert system is 94.5%.

Keywords: expert systems, simulation models, spark maps, precision livestock farming, dairy cows.