Integrated Decision-Making And Optimization Framework For Weapons Under Uncertain Environment

Information technology is achieving the full penetration of military field with its strong vitality,which has a profound impact on weapon system,troop organization,warfare results and decision-making methods.In this context,the traditional decision-making methods,which rely on the commanders' preferences and knowledge,can not adapt to the full-depth,non-linear,full-factor and fast-paced operational requirements of modern warfare.Based on analyzing the characteristics of weapon decision-making problems,the intelligent evolutionary algorithms and modern information technologies are introduced to make accurate analysis and scientific judgments on the complex and large-scale decision-making problems with multiple targets.Therefore,it will be a great accommodation to military commanders in the changing battlefield situation under the conditions of joint operations.In this paper,a series of uncertain factors in weapon decision-making process are discussed,such as the incomplete weapon decision-making data,the uncertain damage probability for targets,and the inaccurate preference information of commanders.To handle these uncertainties,the theories and methods of rough set,military operations research,intelligent evolutionary algorithm and fuzzy multi-criteria group decision-making technology are utilized in this research.Then,the closed-loop optimization framework with feedback effect is constructed for weapon decision-making on the basis of these approaches,including decision rules extraction,applications for single weapon,combinatorial optimization of weapon portfolio,weapon scheme selection,and agile simulation experiment.In this way,it will effectively improve the overall combat effectiveness of weapon systems to provide decision support.The main research contents and achievements of this paper can be summarized as the following five aspects:(1)The closed-loop optimization framework for weapon decision making is established under uncertain environment.In order to satisfy the requirements of making weapon system decisions with effectiveness,accuracy and coordination,this paper proposed a closed-loop optimization mechanism to solve weapon decision-making problems under uncertain setting.According to the logic flow of the optimization framework,the key technologies and approaches consist of four parts: the intelligent two-stage decision model for single weapon application based on rough set and fuzzy reasoning,the combinatorial optimization for weapon system portfolio using evolutionary algorithm,the weapon scheme selection employing multi-criteria group decision-making technology,and agile simulation system for weapon decision-making feedback.(2)The intelligent two-stage decision model is proposed for single weapon application based on rough set and fuzzy reasoning.Aiming at the incomplete decision information in weapon system decision-making process,an intelligent decision model is built for single weapon application with two phases of rule extraction and fuzzy logic.In the first stage,the maximum probabilistic value method is adopted to complete the missing values,then the critical criteria is obtained by using the attribute reduction algorithm on the basis of positive region.Next,the LEM2 rule extraction algorithm is used to learn the fuzzy decision rules for weapon utilization.In the second stage,the Mamdani fuzzy inference system is constructed to improve the automation level of weapon decision making,where the characteristic information about hostile targets are treated as input parameters,and the specific weapon system category is regarded as the output parameter.Ultimately,the final decision result is acquired according to the fuzzy rules extracted in the first stage.(3)The adaptive Memetic algorithm is brought forward to solve the combinatorial optimization and simulation for weapon system portfolio.The objective of weapon system portfolio problem is to determine the appropriate assignment of various weapon units,which maximizes the expected damage of all hostile targets,while satisfying a set of constraints.This paper presented a mixed integer nonlinear optimization model for the weapon system portfolio problem.In order to solve this model,an adaptive Memetic algorithm using crossover and mutation probabilities that are automatically tuned in each generation is proposed.In this framework,the adaptive immune genetic algorithm is adopted as the global search mechanism to accelerate the convergence speed of evolving population.Furthermore,the swap operation-based variable neighborhood search method is embedded to conduct local search and fine-tuning in each iteration procedure.In addition,the VNS/rand/Nr strategy is applied to prevent trapping into the local optima.(4)The intuitionistic fuzzy VIKOR approach is raised to select the optimal weapon scheme within the multi-criteria group decision-making setting.With regard to the weapon scheme selection problem,the important weights of decision makers are always unknown,and the criteria weights may also be completely unknown or partially known.The precise weight for each commander is identified by using the concept of uncertainty degree,which is denoted by the intuitionistic fuzzy entropy of individual decision matrix.For the weapon scheme selection problem under certain constraints,this paper introduced a linear programming model based on the principle of maximum deviation to calculate the criteria weight.For the circumstances in which the criteria weight is completely unknown,this paper suggested a non-linear programming model based on intuitionistic fuzzy score function.Finally,the VIKOR method is combined with GRA in order to achieve the sorting order of alternative weapon schemes.In addition,the existing distributed simulation architectures could not satisfy the requirements of online decision support for weapon application.Therefore,this paper focused on the system architecture of agile simulation,with which the operational plans of weapon application could adjust dynamically.Meanwhile,military experts can make better use of simulation systems to reduce information distortion in agile simulation framework.The design process of agile simulation may absorb the valuable experiences from the typical military analysis systems.This paper explained the agile characteristics from three perspectives of simulation engine,modeling method and data exchange.Two approaches are presented to realize the integration of MSDL and C-BML,with the purpose of improving the scenario reusability and the interoperability between different systems.