There has been studied the influence of gypsum-cement-pozzolana binding agent composition on its key strength properties. The study was aimed on the definition of such consumption of gypsum-cement-pozzolana binding agent components, which could allow maximum ultimate compression strength with minimum consumption of Portland cement provided. As the components of gypsum-cement-pozzolana binding agent, there have been used plaster, Portland cement, and the most wide-spread in the region pozzolana additive. An experiment design was developed to fulfil the work. The research was held in correspondence with the design, and the acquired results made it possible to estimate and analyze descriptive statistics, and expose the influence of components on strength properties of gypsum-cement-pozzolana binding agent. In the course of the analysis it was specified that pozzolana additive consumption factor, and interaction of consumptions of plaster and Portland cement exert a positive influence on strength properties of gypsum-cement-pozzolana binding agent. There have been acquired mathematical models of the influence of gypsum-cement-pozzolana binding agent components consumption on ultimate compression and bending strength of gypsum-cement-pozzolana binding agent. The said model allow forecasting gypsum-cement-pozzolana binding agent composition to have a desired ultimate compression strength. An optimization was carried out and there was acquired a composition of gypsum-cement-pozzolana binding agent with maximum ultimate compression strength with minimum consumption of Portland cement provided.

Keywords: gypsum-cement-pozzolana binding agent, optimization of composition, ultimate strength, experiment design, descriptive statistics, superposed diagram, correlogram, ternar chart