Avaliação de modelos matemáticos para a resolução de job shop problem com a utilização de recursos humanos especialistas em projetos.

Abstract

A project can be defined as a complex system. This requires the use of resources (human, material, technology, etc.), allocated among alternative uses, as a means to achieve specific goals by the presence of constraints of different orders. Traditionally, the techniques of project management attach great emphasis to the definition of attributes associated with the activities (resource estimates, sequencing activities, obtaining estimates of time and cost, schedule and preparation of project budget). Already the project portfolio management aims to establish criteria for the prioritization of projects within a portfolio, but even this deals only incidentally with the trade-off that emerges because of (i) scarcity of experts, (ii) (re) definition of criteria and (iii) assessing the impact of the rescheduling of one or more projects, depending on the (s) objective (s) to be achieved. Regarding human resources, including resource specialists, planning, allocation and prioritization of these resources is made by the portfolio management in a single project to project. This treatment can cause internal strife by using the same resource or even its underuse, and may worsen in software development environments due to the high degree of interdependence, uncertainty and risk of each project. Many software development environments are marked by shortages of Human Resources Specialists - HRS. Although the adoption of smoothing techniques, parallelism and compression, assist in planning and optimization of each project, the interdependence between these demands the integration of management processes. This need is related to the so called job shop problem - JSP. The literature suggests ways to deal with this aspect. A possible solution may be to give through the introduction of optimization methods, involving the minimization of the period or the total cost of the projects in the portfolio subject to precedence constraints or use of RHE, for example. As suggested here, such a method can be developed and deployed based on (i) the technical alternatives and (ii) estimates of resources that include parameters directly related to the construction schedule and budget of the projects in the portfolio. In this context, the objective of this study is to evaluate mathematical models and genetic algorithm optimization and its contributions to solving Job Shop Problem in software development projects using human resources specialists. This research is exploratory, it aims to integrate aspects related to time management, cost management, risk management in view of the need to allocate resources experts in a multi-project environment. As a result, it was possible to identify the advantages and limitations and present a comparative analysis between performance indicators among three mathematical models for solving JSP marked by the presence of skilled human resource allocation in a software development company through a study of case.