摘要
Resource allocation for an equipment development task is a complex process owing to the inherent characteristics,such as large amounts of input resources,numerous sub-tasks,complex network structures,and high degrees of uncertainty.This paper presents an investigation into the influence of resource allocation on the duration and cost of sub-tasks.Mathematical models are constructed for the relationships of the resource allocation quantity with the duration and cost of the sub-tasks.By considering the uncertainties,such as fluctuations in the sub-task duration and cost,rework iterations,and random overlaps,the tasks are simulated for various resource allocation schemes.The shortest duration and the minimum cost of the development task are first formulated as the objective function.Based on a multi-objective particle swarm optimization(MOPSO)algorithm,a multi-objective evolutionary algorithm is constructed to optimize the resource allocation scheme for the development task.Finally,an uninhabited aerial vehicle(UAV)is considered as an example of a development task to test the algorithm,and the optimization results of this method are compared with those based on non-dominated sorting genetic algorithm-II(NSGA-II),non-dominated sorting differential evolution(NSDE)and strength pareto evolutionary algorithm-II(SPEA-II).The proposed method is verified for its scientific approach and effectiveness.The case study shows that the optimization of the resource allocation can greatly aid in shortening the duration of the development task and reducing its cost effectively.
Resource allocation for an equipment development task is a complex process owing to the inherent characteristics, such as large amounts of input resources, numerous sub-tasks, complex network structures, and high degrees of uncertainty. This paper presents an investigation into the influence of resource allocation on the duration and cost of sub-tasks. Mathematical models are constructed for the relationships of the resource allocation quantity with the duration and cost of the sub-tasks. By considering the uncertainties, such as fluctuations in the sub-task duration and cost,rework iterations, and random overlaps, the tasks are simulated for various resource allocation schemes. The shortest duration and the minimum cost of the development task are first formulated as the objective function. Based on a multi-objective particle swarm optimization(MOPSO) algorithm, a multi-objective evolutionary algorithm is constructed to optimize the resource allocation scheme for the development task. Finally, an uninhabited aerial vehicle(UAV) is considered as an example of a development task to test the algorithm, and the optimization results of this method are compared with those based on non-dominated sorting genetic algorithm-II(NSGA-II), non-dominated sorting differential evolution(NSDE) and strength pareto evolutionary algorithm-II(SPEA-II).The proposed method is verified for its scientific approach and effectiveness. The case study shows that the optimization of the resource allocation can greatly aid in shortening the duration of the development task and reducing its cost effectively.
基金
supported by the National Natural Science Foundation of China(71690233)
