Abstract
To solve the concrete mixture design problem, engineers have traditionally relied on guidelines such as those from ACI, and a conservative, labor-intensive, time-consuming, and costly trial-and-error approach that neglects cost or environmental impact of the mixture in the design procedure. In this paper, the concrete mixture design problem is solved through adroit integration of a nonlinear optimization algorithm (OA) and a computational intelligence-based classification algorithm (CA) used as a virtual lab to predict whether desired constraints are satisfied in each iteration or not. The model is tested using previously collected data, three OAs, and three CAs. The outcome of this research is an entirely new paradigm and methodology for concrete mixture design for the twenty-first century. The most cost-effective solutions are achieved by the combination of neural dynamics model of Adeli and Park and enhanced probabilistic neural networks. The cost savings for large-scale concrete projects can be in the millions of dollars.