Study On The Comparison Of Buffer Size Determination Methods Considering Activity Correlation

As the uncertainty of the project gradually increases and the competition between companies becomes increasingly fierce,the critical chain project management method as a new method of project resource scheduling has received extensive attention in academic research and production practice.The buffer determination and setting issues are the core of critical chain project management.The proper buffer size can effectively deal with the impact of uncertainties in the project.At present,the main buffer size determination methods don't consider the correlation between activities.However,in the project practice,due to the influence of logical relations or resource constraints,there is a certain correlation between activities,and the activity correlation directly affects the implementation of the project and the performance of the project.Therefore,it is necessary to study on the comparison of buffer size determination methods considering the correlation between activities.Using the evaluation index system of the buffer size determination method,evaluate the four widely used buffer size determination methods: cut-paste method,root square error method,APRT method,and APD method.Taking into account that there are fewer evaluation indexes,and the index weights are easy to obtain,use the simple additive weighting method as the evaluation method.The entropy method,which has strong objectivity and is less affected by subjective human factors,is used to assign weight to each evaluation index.The acquisition of the evaluation index value is divided into two parts.First,a critical chain project scheduling plan based on the four buffer size determination methods of the cut-paste method,root square error method,APRT method,and APD method is generated.With the critical chain duration plus the project buffer size,the project planning period was evaluated.Then,by simulating the execution process of the project,Monte Carlo simulation technique was used to obtain the other two evaluation indexs: the probability of project completion within the project planning duration and the time of the simulating duration of the project beyond the planned duration.By analyzing the reasons for the correlation between activities,it is clear that resource constraints are one of the main reasons for the correlation between activities.The activity correlation studied in this paper is the activity correlation caused by resource constraints.From the perspective of activity correlation,projects are divided into two categories according to the number of time periods in the project that generate resource conflicts.They are projects where multiple activities have activity correlation within the same time period and multiple activities have activity correlation within multiple time periods.In the process of Monte Carlo simulation experiments,projects are randomly generated,and the type of project to which the randomly generated project belongs is determined based on the number of time periods in which resource conflicts occur.For each randomly generated project,three levels of uncertainty are low,medium,and high.Under the simulation of the project execution,the evaluation indexes of the four buffer size determination methods are calculated.After the weights are determined using the entropy method,the simple additive weighting method is used to calculate the comprehensive evaluation index values of each buffer size determination method,and the applicable situation of the four buffer size determination methods are obtained: in projects where multiple activities have activity correlation within the same time period,the root square error method works best at lower uncertainty levels,the root square error method works best at moderate uncertainty levels,and the cut-paste method works best at higher levels of uncertainty.For projects where multiple activities have activity correlation within multiple time periods,the root square error method and the APD method work best at lower uncertainty levels,the APRT method works best at moderate uncertainty levels,and the APRT method works best at higher levels of uncertainty.Finally,a specific example is used to verify the comparison conclusion of the buffer size determination methods obtained in this paper.Through the simulation of the project execution process,the correctness of the conclusion is proved.This paper provides the decision-making basis to choose the more effective buffer size determination method for the users of the critical chain project management method.