Construction time and cost optimization using A.I. and statistical methods, through Bayes-Point Machines

Scope of the research is the definition of a methodology, composed by innovative tools and techniques, able to optimise time and costs of building activities in facilities that, despite involved in working activities like refurbishments, should guarantee the continuity of activities related to the assigned use of the building itself. Indeed, the current need is to intervene on existing building, adapting performances expressed both regarding to currents laws and building users’ satisfaction rates. Thus, in order to guarantee the quality of the building process, is necessary to know as deep as possible timetables, costs and related bearing of these on the fruitful usability of building and the viability of forecasted working activities. Therefore, in the current practice, to define these values it is widespread the use of standardizes quantity take-off for estimating cost and, on the other hand, Gantt diagram to manage the required time for each activity, mainly relying on the implicit knowledge of project manager. For these reasons, this paper describes a methodology based on agents-simulations, able to support designers in the definition of the construction project strategy, predicting in advance – and relying on more objective basis – the more productive working-phases sequence, based on the full occupancy of available working areas. The proposed methodology starts from the hierarchical breakdown of technical elements of the building, formalizing the Working Breakdown Structure (WBS) and the related link of this along with the locations where activities will be made, as defined in a Location Based Management (LBM) approach. After this interaction, the simulation result will define the duration and the spaces of working areas, defining more reliable cost because they are linked to the effective use of resources, in a defined times and spaces. Results, experimentally verified on real case-study, shows how is possible a reduction of average times required to building working activities, mitigating the impact of these on the possibility to continue the use of facilities and, furthermore, optimize the available resources. Future developments oh this research are connected to the likelihood to extend these techniques toward the facility-management ordinary activities, for the sake to automatize the supply-chain and improve the quality, expressed in terms of reducing time-required to intervene and the interruption of the service provided by the building involved in this process.