Abstract

Concrete is a composite construction material made primarily with aggregate, cement, and water. Other\ncementitious material like fly ash etc. are used which makes it a highly complex material and modeling prediction of concrete\nstrength is a difficult task. Modeling of compressive strength is done using various tools as Artificial Neural Network,\nGenetic Programming etc. In this work an effort is made to predict 28 day compressive strength of concrete using Genetic\nprogramming (GP). GPTIPS, an open source MATLAB based software platform for symbolic data mining (SDM) was used\nfor model development of four models with two sets of data. The multigene genetic programming (MGGP) technique models\nthe compressive strength of concrete by integrating the capabilities of standard genetic programming and classical regression.\nData in Set1 consists of mix design parameters and water absorption parameters as inputs and Set 2 consists of mix design\nparameters and properties of concrete as inputs for conventional concrete and Set 3 consists of non-dimensional parameters\nrelated to recycled aggregate concrete as input parameters. The models were developed in the form of equations using\nGenetic Programming and it was seen that some important input parameters are excluded in the equations. However the\nperformances of these models are noteworthy. A question thus arises about the exclusion of the parameters from the\nequations and its explanation through the theoretical angle.