MODELLING THE NETWORK GRAPH TRAVERSAL PROCESS FOR SOLVING SHORTTERM IT PROJECT PLANNING TASK

Abstract

Introduction. The practice of using network graphs is known for their effectiveness in project management for long-term planning and assessment of the project's current status. The network graph allows you to calculate the main indicators for finding the critical path of work, as well as time reserves for each work separately. This method has proven itself in managing large-scale projects, but the effectiveness of its application for short-term software development planning in  dynamic conditions (conditions of constant changes in content, human and time resources) has not been sufficiently studied. In addition, the difficulty of using standard project management approaches to managing the process of developing specialised software in a dynamic environment (conditions of constant change) leads to the urgent task of researching new, scientifically based methods and approaches for estimating the duration of specialised software development, which will be adapted to the specifics of work in dynamic conditions with a significant degree of uncertainty and resource constraints (human and time resources).
Purpose. Simulating the processes of network graph traversal to calculate the main parameters of the network model will allow for quickly and accurately assessing the duration of project work in the context of dynamic resources and work content.
Methods. To achieve this goal, we used the simulation modelling methods of discrete systems using a Petri net. Given that the object of modelling is the process of traversing a network graph, the conceptual framework of graph theory
is used to describe its key properties. 
Results. Unified models of network graph traversal processes of varying complexity and configuration have been obtained, which allows solving the problems of short-term planning of projects for the development of safety-oriented services in dynamic conditions by calculating the early and late times of events and obtaining characteristics of the duration of project work based on them. The obtained models made it possible to build algorithms for traversing the network graph to estimate the duration of project work with subsequent automation of short-term planning processes, subject to any changes to the content and scope of individual project sprints or the composition of the development team.
Conclusions. A unified model of network graph traversal processes was obtained through simulation modelling using the conceptual apparatus of Petri nets. It consists of the cyclic procedure for establishing a connection between the previous
and current events and allows its adaptive application regardless of the complexity and configuration of network graphs. Through algorithmic reproduction of the developed simulation models, a clear procedure for calculating the early and late deadlines for events has been built, which allows for the automation of these processes to solve the problems of short-term planning of specialised software development projects.