Abordagem evolutiva multiobjetivo para o sequenciamento de tarefas com precedência: um estudo de caso multidisciplinar em mineração

Abstract

This dissertation deals with the problem of task assignment to agents, in the context of project management. An algorithm is developed, such that each task is assigned to an agent who is chosen form a pre-defined set, being executed in a specific moment, complying with precedence relations between tasks. Two objective functions are considered: the project cost and the execution deadline. The multiobjective optimizationalgorithm NSGA-II is employed for solving the resulting multiobjective optimization problem. The proposed algorithm was employed in the study of a practical case, related to the process of planning and management of a feasibility study in mining engineering. Results on the mechanics discipline are compared, concerning the cost and time distribution, task interlocking and capacity of the agents to perform the tasks. The impact of the proposed methodology in project management is briefly discussed.