Recommending the admixture portfolio(s) for high-performance concrete via data-driven decision-making

Admixture portfolio prioritisation and selection for concrete materials has become challenging due to the numerous cost-effective, carbon-exhausted cement replacement options involved in the composites, and the increasing number of items considered for testing each specimen. As the current experience-based process for recommending high-performance concrete (HPC) may not fulfil consulting requirements, we propose implementing a data-driven decision-making (DDDM) process in a materials laboratory to assist experts in providing practitioners with safe and high-quality recommendations for ordering ready-mix HPC for grouting or prefabrication. Two hybrid MADM models, AHP-TOPSIS and AHP-SAW, were constructed, each of which incorporates computerised experimental and expert opinion data to identify the most suitable portfolios (i.e. multi-alternative screening). A rigorous process involving numerical and statistical methods was employed to cross-validate the results obtained by these models individually, and subsequent analysis using kernel set theory showed agreement on three portfolios to be recommended (out of 12). As few studies have addressed the decision-making issues relating to concrete admixtures beyond existing predictive or data analytical works, this study aims to bridge the current research gaps by offering a computerised solution using the hybridised MADM modelling approach, which realises the theory of DDDM.