Optimized High Performance Concrete Mix Proportioning Through JAYA Algorithm

In this paper an interdisciplinary research related to optimization of engineering design which is directed towards sustainability of materials is presented. The case in point is optimization of high performance concrete mixes. The recent bio-inspired algorithm, namely, Jaya algorithm has been implemented for the same. The development of models comprised of two steps, a sizable data of 500 mix designs gotten from standard publications and reported experimental results by researches were preprocessed and legitimate and befitting data sets numbering around 450 were retained for model development. Further, the data is partitioned in to five strength ranges. For each strength range, the lower limit and upper limits of variables and rational ratios of weights were mined from the data after its preprocessing. The algorithm performed very well just for 100-150 iterations. The mix proportions generated by this algorithm for assorted 28 day’s ranges of strength are highly acceptable and align closely with the practical values found in the data. This is reflected by the low mean square error. The mean square error was found to be 10 %-12% for cement, 7% - 9% for fly ash, and 2.1% - 4.0 % for water. Further, a comprehensive comparison of the results obtained in the previous studies of the author with other algorithms namely, Honey Bee Optimization (HBO), Ant Lion Algorithm(ALO), Particle swarm Optimization(PSO), and GA-Elitism Based models (EGA), is also made and presented. The quantities of ingredients of concrete generated by these algorithms are almost close. The differences in mean square errors noticed were marginal though. The output of Jaya algorithm is found to be very close to those generated by ALO.