The benefits of evaluating a new technology in a whole-farm context using a linear programming framework are well known. Linear programming allows the joint evaluation of concurrent farm activities, while considering the costs and returns of all enterprises and any resource adjustments imposed by adoption of the technology. This Report provides a rationale for and description of a whole-farm linear programming model that can be used for the economic evaluation of new technologies that are applicable to beef/sheep grazing farms typical of the Northern Tablelands of New South Wales. In this farming system, the whole-farm focus incorporates various aspects of the pasture base, resource constraints and sheep and cattle interactions. An overview of economic tools that are available to assess technologies at the farm level is provided first, listing some of the major benefits and limitations of each of these various techniques. A representative farm for the selected farming system is then developed and a whole-farm linear program based on this representative farm is described in some detail. A series of modelling experiments is undertaken to examine variations of the base model and their impact on the resulting technology evaluation. An example technology, involving the genetic improvement of beef cattle for improved feed efficiency (NFE), is evaluated. The optimal farm plan for a "typical" (single) year is generated, given the objective of maximising farm total gross margin. Three enterprises are selected: 1,108 first-cross ewes, 1,732 Merino wethers and a beef herd of 127 cows producing 18 month old heavy feeder steers (HFS) at 448kg liveweight and excess heifers sold as 9 month old weaners. For this farm plan, the annual operating budget shows a total gross margin for the farm of $86,191. The optimal farm plan for the representative farm is found to be sensitive to relatively small changes in input or output prices and production parameters. Only small improvements in a number of the individual enterprise gross margins would result in them displacing the currently selected enterprises. These results suggest relatively similar profitability levels between these sheep and beef enterprises. This would be anticipated given that all the enterprises described in this report were identified by local experts as being common in the Northern Tablelands. Further, the relatively small differences in enterprise profitability when viewed in a whole farm context also reflect the similar resources that each of the enterprises require, making them readily substitutable. For new technologies that have dynamic attributes, measuring the cashflow over time becomes important. Genetic traits in ruminants that have long biological lags are such technologies. This means that a single-year equilibrium model will be unable to effectively measure the costs of introducing the new technology over time. In the case of the NFE technology in beef cattle, any herd expansion that is possible as a result of the trait is measured by the opportunity cost of heifer sales forgone that are instead retained to increase the breeding herd. These herd dynamics can be represented explicitly within a multi-period version of a whole-farm LP model. The NFE cow enterprise is offered to the model, with the initial sheep enterprises set the same as the base case (1,108 prime lamb producing ewes, 1,732 19-micron Merino wethers). The model again selects 127 HFS producing cows in the first year, but the new optimal farm plan is to invest in the new technology by purchasing NFE-superior bulls in successive years and expanding the cow herd while concurrently decreasing the scale of the Merino wether enterprise. Substitution of Merino wethers for NFE cows occurs up to year 12 after which additional breeding cows are possible from their increasing net feed efficiency alone. There is an increase in cow numbers of 12.6 per cent by year 25, which equates to an improvement in the NPV per breeding cow per year over the base herd of $5.02, using a 5 per cent discount rate. Other experiments reported include adding constraints for fixed costs, family drawings and an overdraft facility; alternate discount rates for the NPV calculations; alternate terminal values for the livestock assets at the end of the simulation period; and a post-optimality risk analysis. This study has highlighted several additional benefits of evaluating a technology in a whole-farm multi-period linear programming framework. First, apart from determining the type and size of the optimal farm enterprise mix and the optimal value of the objective function, whole-farm multi-period linear programming also provides important additional information including shadow costs and prices and constraint slacks, and how they change over time. Shadow costs of activities show how sensitive the optimal farm enterprise mix is to changes in the gross margins of alternate farm activities not included in the current farm plan. The shadow prices for resources indicates how much a farm manager could pay for additional units of a limiting resource, for example, additional labour. Second, in terms of the specific NFE technology examined in this report, it would appear that there may well be regions where such feed efficiencies may be of greater benefit due to particularly large variations in pasture growth patterns throughout the year. The Northern Tablelands with its recognised winter feed deficit may be one such area. This information may be of benefit to researchers in extending the NFE technology to farmers. Third, the deterministic multi-period version of the model highlighted the impact of the inclusion of overhead and capital constraints in the modelling process in determining the potential adoption of a technology by a farm manager. The availability and cost of capital is shown to influence the extent to which the NFE technology may be adopted by an individual farm business. Fourth, from a modelling perspective, the effect of uncertain terminal values and the bearing that they have on measuring the level of adoption of a new technology is an area for further investigation. Finally, the impact of risk was assessed in this study post-optimally by the inclusion of stochastic output prices in the optimal whole farm budgets. This is an area for further research, including the potential of alternate modelling techniques such as MOTAD programming or stochastic dynamic programming. However due to size constraints, such approaches may necessitate trade-offs in terms of the detail of whole-farm models to which they are applied.