Urban Logic
[03]
Urban Simulation
[Desciption]
A multi-component urban simulation of Riyadh city modelling population growth dynamics, residential mobility patterns, and the spatial allocation and optimisation of general education provision. Combines agent-based modelling with AI-driven optimisation to evaluate how demographic change pressures service infrastructure.
[Model Objective]
A multi-component urban simulation of Riyadh city that integrates demographic forecasting, residential mobility, and education facility optimisation into a single modelling framework. Built as a PhD research tool at UCL's Centre for Advanced Spatial Analysis, the model challenges the century-old practice of using fixed global averages to set urban service standards — replacing it with a dynamic, population-sensitive approach. The framework operates at both city and district scales, and is designed to be generic enough to optimise the provision of any urban service, not just schools.
PhD Research
2023
Riyadh Urban Growth & Education Provision Model
Know more
The model monitors population growth trajectories, school-age demand by district, stage, and gender, and the optimised supply of education facilities over a 50-year horizon (2020–2070) across Riyadh's 150 districts. Calibrated on Riyadh's 2016 population data, sample district results show that current supply levels already exceed projected demand in established districts, while fast-growing peripheral districts face significant gaps. Sensitivity analysis across fertility, migration, and mobility scenarios reveals how robust a given provision plan is under demographic uncertainty — with school-age population estimates carrying an average error of 11% MAPE across tested scenarios.
[Model Components]
City-level population growth and standards optimisation · District-level ABM with household agents · Residential mobility sub-model (Spatial Interaction / Neural Network / Hybrid) · School-age population synthesis by stage and gender · Two optimisation approaches: area-based (m² per student standard) for strategic city-level planning and land parcel-based (pre-designed building prototypes matched to specific available plots) for district-level precision
[Model Results]
The model monitors population growth trajectories, school-age demand by district, stage, and gender, and the optimised supply of education facilities over a 50-year horizon (2020–2070) across Riyadh's 150 districts. Calibrated on Riyadh's 2016 population data, sample district results show that current supply levels already exceed projected demand in established districts, while fast-growing peripheral districts face significant gaps. Sensitivity analysis across fertility, migration, and mobility scenarios reveals how robust a given provision plan is under demographic uncertainty — with school-age population estimates carrying an average error of 11% MAPE across tested scenarios.
[Model Components]
Agent-Based Modelling · Cohort Component Modelling (CCM) · Spatial Interaction Modelling · Neural Network (Deep Learning) · ARIMA Extrapolation · Local Optimisation · ODD Protocol
[Tags]
["ABM", "AI", "Optimisation", "Education Provision", "Demographic Forecasting", "AnyLogic"]