Research On Construction Process Environmental Simulation And Optimization Under Deep Uncertainty

The building activity is one of the important sources of environment impact.The developing countries around the globe are currently in the stage of urbanization and infrastructure development.A great number of house and infrastructure buildings are being or planning in near future.However,the environmental pollution problem becomes increasing severe around the globe.Optimizing the environmental performance while meeting the building demand is an urgent problem need to be solved at present.During the full life-cycle of buildings,construction is a stage in which environmental pollution is generated concentratively.Therefore,optimizing the environmental performance of the construction decision making and realizing the environmently-friendly construction process are important ways to achieve the green construction industry.Therefore,this paper establishes methods of environmental impact assessment,optimization and decision making,aiming to provide comprehensive and systematic decision-making support for environmentally-friendly construction.The complexity of construction activity is reflected into the interwoven and influencing construction processes,as well as the influence from uncertainty factors on construction performance.It is challenge for current static environmental impact assessment method to handle above environmental performance characteristics.Inspired by process simulation technology,this paper systematically integrates discrete-event simulation and process-based life cycle evaluation(DES-p LCA)to establish an assessment method that can fully reflect the characteristics of the environmental impact of construction.Established environment impact assessment method based on process simulation can take advantages of both the process simulation technology and life-cycle assessment.On the one hand,discrete-event simulation reflects complex process and the influence of uncertainty factors on the construction environmental performance,on the other hand,life-cycle assessment ensures the integrity and accuracy of construction environmental impact.The decision-making of construction projects needs to comprehensively consider the project objectives in multiple dimensions,so the reasonable decision-making of construction environmental optimization needs to ensure the performance of both environmental objective and other important project objectives.In this paper,the construction process environment optimization is abstracted as multi-objective optimization mathematical problem.The construction environment optimization is defined as "cost-time-environment" comprehensive goals to seek optimal solutions in multi-objective or take trade-off in different objectives.In the technology of construction multi-objective optimization,the classical simulation-based optimization framework is the most potential and classical optimization method.However,the "simulation-optimization" has inefficient computing that prevents it from providing timely optimization decision-making feedback for construction.Inspired by the fast developing machine learning technology,this paper embeds the learning model into the "simulation-optimization" framework and establishes an modified model "simulation-learning-optimization".The modified model is proved to be able to improve the computing efficiency of the optimization and provide timely decision support for construction,meanwhile it ensures a similar optization quality.In the progress of construction,the uncertainty in terms of both the value and probability distribution of uncertainty factors makes the construction decision making in environment of deep uncertainty.Traditional uncertainty analysis methods,such as Monte Carlo and sensitivity analysis need to base on probability distribution,therefore,they cannot objectively reflect the uncertainty of probability distribution.So they are not applicable for the practical construction decision-making.In this paper,the deep uncertainty theory is used as the theoretical basis for the uncertainty analysis of construction decisions.It compehensively employes the theory and method of Robust Decision Making,Latin Hypercube Sampling and Patient Rule Induction Method to construct the construction decision-making analysis model under deep uncertainty.This model,in one hand,quantitatively compares the robustness of different construction schemes,on the other hand,identifies the vulnerability scenarios that cause robustic schemes with unacceptable performance using data mining technology.The information is an effective basis for environmentally-friendly decisions under construction deep uncertainty.The theoretical achievements mentioned above are the main innovations of this paper.Finally,the implementation system of DES-p LCA,"simulation-learning-optimization" and deep uncertainty decision analysis method are built based on above theoretical model.They are implemented to a real construction case,the optimal parameter settings of proposed models are obtained,and the validity and applicability of the theoretical model are valided.On the research path of environmentally-friendly construction,this paper proposes environmental simulation assessment,multi-objective optimization and uncertainty decision-making support methods that take accounts of the characteristics of construction environmental pollution generation and construction decision making.Innovative simulation based environmental assessment method is the progress of environmental impact assessment theory and method.The "simulation-learning-optimization" model is an extension of the classical "simulation-optimization" framework.Deep uncertainty decision analysis model promotes the development of uncertainty analysis theory in construction field.The research of this paper is the systematic method that can be used to implement environmentally-friendly construction.