This paper illustrates the use of activity modelling at a detailed scale to determine the teaching space requirements for institutions such as universities or polytechnics. It is argued that traditional methods of predicting the demand for space, on the basis of a number of utilisation factors, are unsatisfactory because no adequate theoretical explanation can be advanced for the choice of one level of utilisation rather than another. By contrast, in this paper we propose two methods of predicting the demand for space that are based on a detailed understanding of the ways in which any pattern of teaching must be organised in time and space.In the first two sections we examine a method which relates, through the use of a timetabling program, the number and type of lectures, seminars, and practicals to the demand for space. In the third section this use of the timetabling program is extended through the introduction of a solution space within which all possible timetable solutions must fall for a particular set of constraints. From the maximum dimensions of this space we show that it is possible to estimate the minimum and maximum length of the teaching week, and the minimum and maxiumum number of rooms of different types and sizes required for any pattern of teaching.In the fourth section we describe a second method of determining the demand for space, which is based not on a measure of the minmum demand but on the most probable distribution of teaching events over the week. The probablistic approach, it is argued, provides a more realistic measure of demand by taking account of factors, such as inefficiency in timetabling or the unwillingness of staff to teach at certain times, which in practice increase the demand for space above the minimum. Finally the paper ends with a comparison between the two methods proposed and the traditional methods of estimating the demand for teaching space.
