BARK--Budget allocation by ranking and knapsack

Budget Allocation by Ranking and Knapsack (BARK) has been developed to solve a particular problem of determining which projects should be selected (from a given array) for implementation, subject to budgetary constraints. This problem is often encountered in the public sector, where the values (in financial terms) of competing projects or programmes are difficult or impossible to quantify, but where the projects may be ranked in terms of their perceived worth or benefit. Provided a reasonable level of agreement on ranking can be achieved, the rankings can be converted to surrogate measures of value, so that the selection of projects then becomes a standard knapsack problem. The paper describes how BARK has been used in a particular case and how the methodology can be extended to situations when the budget is not defined as a single value, but as a range. The methodology may be applied in many areas, for example in the allocation of an R&D budget, and is therefore not confined to the public sector.