| IT project portfolio management is a typical complex large-scale system with a large number of independent individual projects which interact with each other and lead to global behavior which is dynamic,uncertain,difficult to reproduce,or non-repeatable.IT projects generally have a certain amount of intelligence,but for a long time,set by traditional project management theories as inanimate individuals,which leads to large actual and theoretical diviations of the IT project portfolio management or difficult to apply optimization results.This paper adopts the ideas and methods of hierarchical control,swarm intelligence,multi-agent and information entropy to construct the social computing method and model of IT project portfolio in cyber-space.The method aims to abstract,model,simulate and optimize the IT project portfolio management process.Simulation experiments and practices are carried out using actual data from 11 batches in three years.The main achievements and innovations in this paper are summarized as follows:(1)Aiming at IT project portfolio management,a spatial-temporal hierarchical intelligent social computing method called WFR/3M is proposed.The intelligent simulation model of ecosystem(Silkworm)including the IT projects,the project portfolio and their environment is given.The portfolio batch scheduling optimization model(Rearing Farm)is established in which a hierarchical coordination control scheme is applied to achieve synergy control at the tactical level.A quantitative visual simulation platform(Decision-making Room)is provided to managers and the decision-makers to realize the modeling,analysis,optimization,control and management of the IT project portfolio at the tactical level.(2)A logical model of the ecosystem formed by IT project individuals(P-Silkworm),IT project portfolio(P-Silkworms)and the environment is proposed.A bionic agent model of individual IT project based on Generalized Artificial Life is constructed.Based on this,the complex network theory is used to establish the structure model of IT project portfolio in which a measure of structure entropy is proposed to measure the complexity of the network model.Based on these,the characteristics of the structure model,the spontaneous evolution and the reasons for the evolution of the IT project portfolio batch are analyzed.Thus,the internal relationship and evolution results of IT project portfolio can be quantitatively analyzed.(3)A model for the evolution and optimization of IT project portfolio based on Multi-agent system(MAS)is proposed.The agent roles and interaction mechanism and rules with local perception and optimization strategy are designed.The evolution model of project portfolio network is established to form multiple communities with certain cooperation ability automatically.An improved glowworm swarm optimization(GSO)algorithm based on ensemble learning is proposed to optimize the IT project portfolio structure locally.A global optimization algorithm based on information entropy for the efficiency-risk balance is proposed to realize the global structural optimization of the project portfolio implementation.The above methods ultimately achieve the autonomy evolution and optimization of project portfolio planning,which achieve the trade-off of the efficiency and the risk and improves the portfolio implementation success rate.(4)A human-machine interaction decision house model is proposed.From the perspective of man-machine coordination,the comprehensive evaluation,risk monitoring and risk handling of combined projects at the global level are realized by computer aided methods.A method of unsupervised decision tree method for mining expert scoring behaviour is proposed,which can obtain more objective comprehensive evaluation results.A node risk evaluation method based on Self-organizing map(SOM)net and complex network is proposed to identify key projects,and a resource self-coordination method based on crowd-funding mechanism and a man-machine based on case based reasoning are proposed.The negotiation mechanism deals with sudden risk events and implements macro intervention on the whole.The WFR/3M method proposed in this paper combines the interaction,evolution,optimization mechanism and human-computer interaction negotiation mechanism of the individual and portfolio models.The implementation of the whole project portfolio emerges to a certain degree of intelligence guided by this method.The effective control of portfolio risk can be realized and emergencies can be handled effectively. |
PDF Full Text Request