A study in the application of a Monte Carlo simulation farm planning technique

Abstract

The general problem in farm management is to allocate scarce farm resources among production alternatives to maximize an objective which is usually profit. Farm plans are needed which are technically feasible and which maximize the objective.

In recent years the technique of Monte Carlo simulation programming has been added to the list of tools or aids for solving farm management design problems. The concept of Monte Carlo simulation for farm management planning was first put forward by Lindgren and Carlsson. It has been adopted by several people involved in the systems research, including Thompson, Dent, Donaldson, and Webster.

Simulation is a representation or parallel of real life. For example a fertilizer field trial represents a larger area such as a complete soil type. The experimental plot is thus a small scale parallel or simulation of a larger real life situation. A non-physically experimented solution to a farm management problem depends on a model of a farm situation. A model which as accurately as possible represents or simulates, the real world situation is required subject to costs.

Any simulation which contains one or more random elements is usually termed a Monte Carlo simulation. The term comes from Monte Carlo City where random outcomes to gambling games take place.

Monte Carlo programming is essentially a trial-and error process or an experimental approach. The technique produces sub-optimal farm plans by considering simultaneously all possible farm constraints and activities via a trial-and error system. Only a proportion of the plans built in this fashion result in profitable solutions. This system would not be practical if it were not for very high speed computers and a sorting and storing routine which isolates solutions near the optimum. It is quite different to a linear programme which produces an optimal solution in an iterative process.