Designing the structure of a construction project operating system using evolutionary algorithm

The paper discusses the problem of designing a construction project’s organisation structure at the operating level, where tasks and processes are of “complex of operations” type. Such a system includes heterogeneous operating units (crews, teams) of a general contractor and cooperating external ones – subcontractors, that create a temporary organisation. Its structure changes according to the project schedule as the project advances. The author identified the problem of designing a construction project operating system structure from the point of a general contractor, built the system’s model and formalised it mathematically. The contractor selection process (i.e. the selection of the system’s elements) is described as the problem of triple-criteria optimisation of the schedule. The assessment of possible variants of the system’s structure is made according the criteria crucial for the project’s efficiency and the general contractor’s objectives (i.e. minimisation of project duration and cost, and keeping subcontracting to minimum – as the general contractor is assumed to be interested in making full use of their own resources). To solve the problem, a method that uses metaheuristic approach has been worked out. An evolutionary algorithm (using stochastic processes) was adapted for solving the triple-criteria schedule optimisation problem in deterministic conditions. The author developed also a heuristic algorithm to allocate limited resources of variable availability. The solutions are generated by means of an achievement scalarising function, which is based on Tchebycheff utility function. The selection of final solution can be done by analysing the approximation of whole set of non-dominated solutions on the basis of total decision maker’s preferences, or by means of Steuer’s interactive method.