Research On Project Partner Selection Problem Under Uncertain Environment

The rapid development of information technology and economic globalization has promoted vigorous expansion in all walks of life,which has also brought unprecedented development opportunities to various project-oriented business operation modes.At the same time,projects are becoming larger in size and complexity,especially projects above designated size,which makes it hard for single enterprise to cope with project and it is necessary for an enterprise to choose partners in order to successfully complete the projects.There are many uncertainties in project operation environment and course,such as policy,weather,finance and people's behavior and psychology;which enhances the uncertainty and risk to project completion.Therefore,as an important part in project management,selecting partner under uncertain environment becomes crucial.Whether selecting partners successfully or not will directly affect project performance.Projects'customers are their ultimate owners and any project operation serves for customers.How to reduce completion risk,lower cost and maximize customers'satisfaction level is what enterprises pursue in such fierce competition,which concerns enterprises' sustainable development.One of enterprises' aims is to maximize customers'satisfaction.In real life,subprojects' completion times are often uncertain and there is complex precedence relationship among subprojects,which brings great completion deviation to project and thus leads to completion risk.Therefore,customers' satisfaction level is decreased.Based on the context of partner selection in project management,this paper studies how to ensure overall project's completion time due to subprojects'uncertain completion time.The main research work can be summarized as follows:(1)The problem of project partner selection based on Value-at-Risk(VaR)describing risk is researched.Project execution is liable to be affected by uncertainties which leads to subprojects' uncertain completion times and thus brings project's delay risk.VaR is simple in depicting risk with loss value and independent on probability distribution assumption.So VaR is introduced to measure risk here.A nonlinear model is established with the objective of minimizing the completion cost and the constraint of delay risk.Then the model is transformed into a certain model by stochastic programming theory.According to the characteristics of the model,a hybrid ant colony-genetic algorithm is designed based on subproject precedence relationship.Comparisons among enumeration algorithm,ant colony algorithm and max-min ant system are tested on three different scales of examples.The experimental results are analyzed to verify the validity of the algorithm and good adaptability to different scales.Finally,the proposed model is compared with the other two uncertain models.The results show that the proposed project partner selection model based on VaR describing risk is efficient.(2)The problem of project partner selection considering customers' time preference is studied by prospect theory.Under uncertainty environment,people's behavior directly affects decision process.In order to increase customers' satisfaction with project,the problem of project partner selection is investigated considering customers' time preference from the perspective of customer behavior.Based on prospect theory,a nonlinear programming model is established with maximizing the prospect value of project completion time as target and project expect cost and subprojects' precedence as constraints.According to the characteristics of the problem,a hybrid genetic-simulated annealing algorithm is designed to solve the model and then compared with genetic algorithm.The effectiveness of the proposed algorithm is verified by numerical simulations.In the end,the proposed model is compared with the expect value model by parameter analysis.Numerical results show that the proposed model can better depict customers' psychological behavior.(3)The problem of project partner selection considering customers' time and cost preference is presented.Compared with focusing on the deterministic values of project cost and completion time,some customers pay more attention to the gap between the actual value and the expected value of project cost and completion time,preferring larger confidence level of total completion time and cost approaching to their expected values.Prospect theory is used to describe customers' psychological preference to the satisfaction level of completion time and completion cost.The mathematical model of the problem is established with the optimization objective of maximizing the prospect value of the confidence levels of project's completion time and completion cost.A hybrid ant colony-simulated annealing algorithm is designed to solve the model.By comparison with ant colony algorithm,the effectiveness of the proposed algorithm is verified by calculating the examples.And the model considering time and cost preference is compared with the risk-neutral model.The results show that the model considering customers' time and cost preference can more accurately describe customers' psychological behavior and suitable for customers with different risk attitueds.(4)The problem of project partner selection is presented considering customers'time preference and psychological probability.Under uncertain environment,project customers' limited rational behavior not only lies in recognition bias on uncertainty but also in probability distortion,such as amplifing small probability and lessening large probability,which affects decision.Hence,in order to accurately describe and measure customers' satisfaction level and increase enterprises' core competitiveness,cumulative prospect theory is adopted to consider customers' time preference and psychological probability,and thus the model of project partner selection problem is established.According to the characteristics of the problem,an improved ant colony algorithm based on the information guiding on critical path and max-min ant system is designed to solve the model.Then the results are compared with the ones solved by ant colony algorithm.The validity of the proposed algorithm is proved by simulating results and run time.Furthermore,the established model is compared with the expected value model and the prospect model.The results show that the cumulative prospect model can describe customers' psychological behavior more effectively.