Research On Multi-project Scheduling Under Uncertain Environments

As the continuous improvement of customer requirements, project scheduling isalways under uncertain environment, it often causes that the real scheduling schemecan’t be fully implemented in accordance with the benchmark scheduling scheme orneed to re-scheduling the plan. The multi-project scheduling problems underuncertainty environment take consideration of the uncertainty factors in the schedulingenvironment, such as the delay of project duration caused by the weather, the shutdowncaused by equipment failure, and this uncertainty is inevitable. Therefore, themulti-project scheduling under uncertainty environment has great research value andpractical significance. For the multi-project scheduling problem under uncertaintyenvironment, we need to take some measures to restore or to reduce the influence ofuncertain factors on the multi-project scheduling, in order to ensure the multi-projectscheduling complete smoothly.This thesis is supported by the national socal science Foundation project“Theinland open processing trade mode based on the product division (No.11BGL006)”and major decision consulting research of Chongqing“Emergency material reservesand scheduling mechanism of social”. Firstly，this thesis considers non-fixed durationof a multi-project scheduling problem since that the direct influence of the uncertaintyof the multi-project scheduling is reflected in the change of the project duration; On thebasis of activity splitting using non-preemptive execution mode, Secondly，this thesisstudies the activity of uncertain multi-project scheduling problem for low robustness ofmulti-project scheduling under uncertain environment and the status of low resourceutilization; Nextly, this thesis considers multi-project scheduling problem withinterferences in the process of project implementation under the uncertainty factors;finally，In view of practical problems of the deviation between the actual schedulingand benchmark scheduling under the uncertainty, this thesis considers multi-projectproactive scheduling problem under uncertain environment. Research work is asfollows：①The research is carried out considering the low robustness of multi-projectscheduling under the uncertain environment and uncertain project duration problem.Multi-project Scheduling Duration-Robustness Optimization Models is establishedunder the uncertain environment, because of its complexity, the multi-level geneticalgorithm is designed to solve this problem. The study found, the robustness ofmulti-project scheduling which has durations uncertainties increased with the project deadline. The more stable resource use, the higher the robustness of the projectis.②The influence of uncertain factors on the resource utilization of multi-projectscheduling is analyzed, activitys are splited according to the enforceability of a numberof sub-activities, and the non-preemptive execution mode is performed. The influenceof uncertain factors on the multi-project scheduling system is reduced based onimproving resources utilization of multi-project scheduling. Using ProjectDuration-Cost-Quality-Robustness as the optimization goal, the multi-objectiveoptimization model is established based on the utility function and the cloud geneticalgorithm is designed to solve the problem. Effectiveness of the model and algorithmwere verified by a case study. Research shows that, adopting the non-preemptiveoperation mode, not only can realize the overall utility maximization of multi-projectscheduling, but also improve the ability of dealing with uncertain factors ofmulti-project scheduling, and improve the level of resources utilization of multi-projectscheduling in the case of a shared resource.③The effects of uncertain factors on the project scheduling with multi-factordisruptions problems is analyzed, considering the effects of customer satisfaction, costand available resource Its objective is to make reasonable disturbance recovery strategyand minimize disturbance in the project scheduling under the limit of resourceconstraints and project duration. The model of multi-project scheduling model withmulti-disturbances is established, and the dual-population genetic algorithm isdesigned to solve this problem. Through the analysis, it is found that for multi-projectscheduling problem which faces many interference, using local heavy schedulingscheme can conform to the actual needs of the project, and can more quickly makingplan to restore the interference events affect scheduling system.④The problems of the deviation between the actual scheduling and benchmarkscheduling caused by uncertain factors is analyzed, and the proactive schedulingstrategy is proposed to analyze the disturbing incident based on past data. We alsomake the benchmark scheduling plan is formulationed to predict the possibleinterference and recovery strategy is established. Our goal is to minimize the delay ofthe duration and cost losses under an uncertain environment. The multi-projectproactive scheduling model under uncertainty environment is established and geneticsimulated annealing algorithm is designed to solve this problem. Finally the examplesgive the conclusion and demonstrate the effectiveness of the proposed algorithm. It isfound that, with the increase of project cost the sensitivity from construction period tothe disturbance events drops. When the coefficient of cost factor weight increases, theinfluence of disturbance events on the optimization goal shows a trend of decrease.In this thesis, multi-project scheduling problem under uncertain environment isresearched, but more difficult to uncertain factors identify and describe of disturbanceevents, there is other aspects to be explored in futher research.