A Dynamic Knowledge Model of Project Time-Cost Analysis Based on Trend Modelling

Abstract

This paper investigates the application of trend quantifiers of project time-cost analysis as a tool for decision-making support in the project management. Practical project management-related problems are solved under information shortages. It means that methods of statistical analysis cannot be easily used as they are based on the law of large numbers of observations. Numbers are information intensive quantifiers. The least information intensive quantifier is a trend; its values are increasing, constant, decreasing. If a derivative cannot be quantified by a trend, then nothing is known and therefore nothing can be analyzed/predicted. For this reason, the trend model M was created. The model M is based on a degraded set of differential equations or heuristics. A trend analysis of the model M is an evaluation of the relevant discrete set of solutions/scenarios S. A trend reconstruction is an evaluation of the model M if a (sub)set of scenarios S is given. The paper studies linear reconstruction, i.e. the model M is a set of linear differential equations. The trend reconstruction is partially reverse process to trend analysis. A case study has 7 variables (e.g. Project duration, Direct personnel costs, Indirect personal costs etc.) and the reconstructed set of linear differential equations has 7 equations. The set of 243 scenarios is obtained if this reconstructed set of trend linear equations is solved. Any future or past behavior of the model M can be described by a sequence of obtained scenarios.