Optimering av soltillgång i stadsplanering : En jämförelse mellan genetisk algoritm och Monte Carlo-metoden

In this study, two stochastic optimization methods are compared, Monte Carlo and genetic algorithm, by optimizing solar access in urban planning. The analysis is based on six solar-related performance indicators in relation to seven building parameters, from a dataset for the Stockholm area. The methods are evaluated based on performance, robustness, and computation time. Monte Carlo showed high robustness and short computation time, while the genetic algorithm achieved higher performance but took longer and had lower robustness. The results suggest that method choice should depend on purpose, where Monte Carlo suits broad analyses, while genetic algorithms are better for optimizing specific solutions. The study highlights the influence of building geometry on solar access and recommends incorporating more sustainability indicators in future research.