INTRODUCI'ION
The search for solutions to unintended changes to agricultural systems is a prior;ty for many at the moment. One aspect of this search is the search for management strategies that are more responsive (rather than less responsive) to the current and expected conditions of the system. This search includes the identificaticn of actual strategies and the development of new strategies.
The evaluation of actual and new management strategies is often based on modeiling or simulation. In building these models a distinction needs to be made between exogenous and endogenous determinants of changes in the system. At the one extreme. 11 determinants of change are exogenous. In this case the modelier is in fact determining the ;>utcome of changes in the system by the choice of-data. The model provides structure and executes !he required calculations. At the othpr extreme ail determinants of change are endogenouu. Neither of these two extremes are of much use in :he evaluation of actual and new management strategies. The first because they add little new understanding; the second because they are likely to be too large and too complex.
The model to evaluate alternative managr:ment strategies needs only to contain a sufficient subset of variables that can adequately represent the alternative management strategies. The identification of this subset ofvariables and the evaiuaiion of alternative managemen: strategies is usually more productive when embedded into the provi.:ion of decision support.
The paper outlines attemP,s to identify such a subset of variables and to evaluate such management strategies for the number of breeders to retain on an extensive beef cattle property; aa; concludes that to date this effort has only had limited success. While the example is specific to a beef cattle property, the type of problem is more general. This presentation is seen as a way to stimulate the search for a more appropriate subset of variat+s and more appropriate ways ;o represent and to evaluate mar;agement alternatives that are more respoasivt:
(rath<r than less responsive) to the current and expected conditions of the system A BEEF CATTLE PROPERTY In terms of decision support, an extensive beef cattle property has characteristics that make it rather different to other properties. It can be modelled as follows.
Topographically it consists of a large piece of land (say 53 000 ha) which is partitioned into areas 037X4754/92/ LOO ti 1992 -Elxvicr Scicncc Publkhcrs B.V.