During the past few decades, many global optimisation and multi-objective evolutionary algorithms (MOEAs) have been developed. Those algorithms have shown very useful in enabling system design automation and globally accurate modelling. However, there is a lack of systematic benchmark measures that may be used to assess the merit and performance of these algorithms [1],[2],[3],[7],[8]. Such benchmarks should be consistent with those used in measuring conventional optimisation algorithms, should be simple to use and should result in little program overhead. This paper attempts to formalise, and to promote discussions on, this issue. In this paper, benchmarks in terms of (i) optimality; (ii) solution spread measure; (iii) optimiser overhead will be presented.