Discrete-Voxel Optimization of Parametric Models of Urban Infrastructure

Abstract

This paper develops and validates a discrete-voxel modelling method for the quantitative assessment of microdistrict changes when a new residential building is introduced into existing urban infrastructure. Urban space is represented as a regular 3D grid with a discrete functional state for each cell. Using the integration of a 108-apartment building into an established neighbourhood, we simulate changes in volumetric and planar building density, insolation regime and solar-gain balance, transport accessibility, and green-space provision. The model is implemented in MagicaVoxel: the base grid step is 3 m, and voxel states encode land-use types (built-up, transport, green areas, etc.), enabling algorithmic computation of areas, volumes, and distances. Density is evaluated by global and local measures; insolation by an energy-balance calculation for a representative winter day; accessibility by minimum travel time to transport hubs. For a 10-storey option, volumetric density increases by 1.8%, insolation potential decreases by ~11%, average accessibility time rises by 8–9%, and green-space indicators drop by 2–3% while remaining within regulatory limits. A height variation study (10–12–15 storeys) identifies threshold values where insolation and transport performance deteriorate markedly, providing a basis for selecting acceptable design ranges. Due to the grid-based structure, computations are scalable and well suited to early-stage decision support and rapid comparison of alternatives. Overall, the approach offers a transparent multi-criteria workflow that can be extended to automated search of parameter sets under normative constraints.