Kudin O. Mathematical models for estimating the complexity of work in time management of projects for the development of design documentation of ships

Dissertation for the Degree of Candidate of Technical Sciences. Specialty 05.13.22 – Project and Program Management. – Admiral Makarov National University of Shipbuilding, Ministry of Education and Science of Ukraine, Mykolaiv, 2021. Existing methods and models of project management, in particular methods and models of time management of projects of development of design documentation of vessels, do not always allow to execute the specified projects in the planned time. The presence of cases exceeding the planned project deadlines requires the creation and use of the latest methods and models of shipbuilding project management, including methods and models of time management of such projects. The dissertation solves the current scientific and applied problem of increasing the reliability of estimating the effort on the development of design documentation of vessels under the influence of random unforeseen factors by building mathematical models of effort based on bivariate Johnson normalizing transformation of the SB family. Such mathematical models allow to increase reliability of estimation of effort of works at time management in projects of development of design documentation of vessels. The advantage of the proposed solution is that the assessment of the complexity of the work is based on statistics of completed projects of design documentation development in contrast to existing methods and models of effort assessment, based on the own experience of project managers. As a result of the research, a non-Gaussian probabilistic model of labor intensity for the development of design documentation of vessels hull sections and mass of these sections was built for the first time on the basis of the probability density function of the bivariate Johnson distribution of the SB family. data compared to existing models based on the densities of univariate distributions. The equation of the transformed prediction ellipse for the bivariate non-Gaussian effort vector has been improved, which allows to determine the presence of bivariate emissions in such non-Gaussian data sets without assuming their normal distribution and, further, increase the reliability of effort estimation. The nonlinear regression equation and the limits of its confidence interval and prediction interval are constructed on the basis of bivariate Johnson normalizing transformation of the SB family, which allows to increase the reliability of estimating the sample mean effort in comparison with the corresponding equations constructed by univariate normalizing transformations. The use of these estimates of the sample average effort of work allows, in turn, to increase the reliability of the estimation of the average time intervals of the schedule of the project of development of design documentation of vessels. A nonlinear regression model is constructed to estimate the complexity of work on the development of design documentation of vessels hull sections based on the bivariate Johnson normalizing transformation of the SB family, which reduces the width of nonlinear regression prediction intervals compared to univariate normalization transformations. A method for estimating the average effort in time management of development projects design documentation of vessels using a nonlinear regression model of effort and confidence intervals based on the bivariate Johnson normalizing transformation of the SB family is developed and an example of its use is given. A method for effort intensity as a random variable using a nonlinear regression model of labor intensity and prediction intervals based on a bivariate Johnson normalizing transformation of the SB family is developed and an example of its use is given.