Globally accepted guidelines for land use allocation in Riyadh, Saudi Arabia have been
based on an outmoded practice that was created over a century ago. This approach is
based on a mix of predetermined population densities, walking distances, and per person
area ratios. The latter criterion is essentially based on a worldwide average for facility
areas and user numbers. The fundamental criticism levelled at such practices is their
insensitivity to population trends and limited land resources. In this context, this research
is aimed at updating common practice in the light of population growth and residential
mobility projections at the city and district levels. The models introduced aim to provide
comprehensive and adaptable simulation tools for optimising any type of land use
provision standard over a specified time period. The simulation environment makes use
of an agent-based framework that adapts and integrates a number of well-known
methodologies, including Cohort Component Modelling (CCM) for population
projection, Spatial Interaction (SI) modelling for residential mobility, and
AutoRegressive Integrated Moving Average (ARIMA) for various ratio extrapolation.
Additionally, new hybrid concepts and approaches have been evaluated, including a
household based CCM and the use of Neural Network algorithms (NN) to forecast
residential mobility. The case study focuses on Saudi populations in Riyadh, Saudi
Arabia where the three general education stages at elementary, middle, and secondary
levels were optimised for both genders. Moreover, the optimisation time horizon spans
50 years, from 2020 to 2070 while the focus of research at the city level optimises the
conventional ratio of area per student based on the present stock of education allocated
land and a land consumption ratio defined for every five years. The district level
optimisation, on the other hand, balances the demand and supply of education over 50
years by utilising the Ministry of Education’s (MOE) predesigned school prototypes. The
research findings demonstrate the feasibility of developing a tool for optimising land use
guidelines that is capable of producing acceptable outcomes while being sensitive to
demographic change and land resource availability.