Computational methods for multi-criteria decision support in urban planning

Urbanization and climate change induce many challenges for today’s cities. These challenges have to be faced by urban planners during complex planning processes that involve many stakeholders. Interactive computational approaches can provide assistance during these processes. The goal of this thesis is the development and demonstration of a computational framework that shows the impact of decisions taken during different phases and on different scales of an urban development project A key aspect of the proposed approach is its ability to quickly generate and efficiently handle a large number of alternatives.

Urban systems consist of many nested, partly autonomous elements. Moreover, they are embedded into larger systems and evolve constantly. The resulting unknowabilities comprise the uncertainty of current states and the unpredictability of future states of urban systems. The following computational methods are employed to address these aspects: (i) Multi-parametric programming allows to capture decision spaces of several actors with inherent trade-offs and tipping points. (ii) A corresponding model integrates five domains and four spatial scales. (iii) Multiple linear regression models serve to reduce uncertainty of input parameters. (iv) A data model allows to manage the large number of explored urban scenarios.

The developed system is demonstrated based on three case studies for which specific questions are stated. Starting with a greenfield development project in Europe, a first set of questions addresses the relations between the built density, the sustainability of the energy supply, the distribution of buildings, and the costs of different actors. A sensitivity analysis identifies the impact of changing energy prices on the preferred energy and urban systems. The results reveal tipping points regarding for example the preference for a decentralized or a centralized energy system. Changing from energy issues to livability aspects, relations between the built density, the share of parks, and the view on a landmark are quantified. The other two case studies imply a move from Europe to Asia, and from new developments to existing neighborhoods, respectively. From this result different boundary conditions and questions concerning e.g. maximum achievable densities, energy autonomy, and cost-effective building refurbishments, accounting for heritage protection. A last set of questions addresses the influence of the considered spatial range on the outcomes: increasing the range can reveal synergies leading to overall better solutions, such as the allocation of refurbishment subsidies to neighborhoods where more energy savings can be achieved for less investments. The addressed practical questions demonstrate the potential of the developed system to explore more thoroughly and quickly the decision space in urban planning.