Multi-criteria building energy performance benchmarking through variable clustering based compromise TOPSIS with objective entropy weighting

Enabling robust energy benchmarking to reliably locate performance inefficiency for upgrading is critical to the success of building retrofitting programs in building sector. Multi-criteria benchmarking is emerging as a more rational option over the traditional single-angle method to assess building performance which is fundamentally of multi-factor nature. Particularly, with its easier concept, the compromise Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) based multi-angle benchmarking appears attractive. Nevertheless, existing TOPSIS based procedures tend to ignore the common issue of multicollinearity trap which could result in misleading decisions. Meanwhile, variable clustering renders an empirical alternative for handling multicollinearity with high traceability. Combining with information-oriented Shannon entropy, this paper develops an iterative Clustering around Latent Variables (CLV) based objective entropy weighted TOPSIS approach for benchmarking building energy performance in a multi-factor manner. It essentially integrates the benefits of variable clustering to address multicollinearity with information theory for objective weighting on decision attributes in order to pursue TOPSIS benchmarking accuracy. A 324-dwelling case shows the robustness of the constructed procedure in a temporally dynamic context.