武器装备体系组合规划的高维多目标优化决策

针对武器装备体系组合规划问题中存在多类相互冲突的高维多目标问题(目标数 ≥ 5),提出了一种三阶段的集成优化决策方法. 首先运用目的规划技术将高维多目标问题转换为一般多目标优化模型(目标数 ≤ 3); 然后提出一种多目标差分进化算法,用于搜索属于决策者关心区间的非劣解集; 最后提出基于预测优化的理想点算法,可生成精确满足决策者偏好的最佳折衷解. 通过某侦察装备体系组合规划示例,证明了各算法模块的优势和该方法的整体有效性,可为武器装备发展和顶层规划提供决策支持.

Many-objective optimization and decision-making for portfolio planning of armament system of systems

In order to attack the many-objective problem in armament system of systems (ASoS) portfolio planning (the numbers of objectives ≥ 5), an integrated optimization and decision-making approach with three-stage is proposed. This approach first applies goal programming technique to convert this problem formulations with large-dimensional objectives into a normal multi-objective optimization model (the numbers of objectives ≤ 3), a multi-objective differential evolution algorithm then is developed to solve the converted model and obtain the Pareto set, a forecasting and optimization based ideal solution algorithm finally is proposed to obtain the best fit compromised solution from the Pareto set. A case of portfolio planning for reconnaissance ASoS illustrates the effectiveness of each module and the integral of the proposed approach. This approach can provide the decision-making support for armaments development and top-level planning.