WHAT THE EVIDENCE SAYS
Researchers optimized local-climate-zone layouts in eight Greater Kuala Lumpur districts using machine-learning surrogates and evolutionary search. Modeled surface-heat-intensity reductions averaged roughly 2.85–3.03 °C under population and GDP constraints, depending on the algorithm.
WHY IT MATTERS / OUR READ
Landscape planners can use predictive heat models to compare spatial configurations, but simulation outputs should remain hypotheses until design feasibility and measured thermal performance are tested.
CAVEATS / UNCERTAINTY
Surrogate optimization—not a built intervention, physical climate-model validation or measured air-temperature benefit; implementation costs and tradeoffs were not fully evaluated.
SOURCE
Yasser et al. · Applied Spatial Analysis and Policy (3 Nov 2026)
View original evidence ↗