МЕТОДИ ТА ЗАСОБИ МОДЕЛЮВАННЯ ПЕТЛІ БОЙДА У ВІЙСЬКОВИХ ЗАСТОСУВАННЯХ З ВИКОРИСТАННЯМ ОНТОЛОГІЧНОГО ПІДХОДУ
Oborska O. V. Methods and means of Boyd loop modeling in military applications by using ontological approach. – Manuscript.
Thesis for a candidate's degree in technical sciences in speciality 01.05.03 – «Mathematical and software support of computer machines and systems». – Lviv Polytechnic National University, Ministry of Education and Science of Ukraine, Lviv, 2016.
In the thesis scientific and practical task of methods and means development for building decision support systems in competitive environment by using ontological approach is solved. The efficiency increase of such systems through the use of developed mathematical tools and software supply, based on the use of ontology in these systems, the adaptation of ontology to the specificity of domain problems is achieved. During the research the following scientific and practical results are obtained.
The development expediency of mathematical models, methods and tools for decision support in a competitive environment with the use of the ontological approach to the subject areas in which knowledge is clearly regulated are justified. Such a subject area is the military sphere.
Boyd loop model based on Moore automaton is developed. The states of the automaton are the Boyd loop stages as well as processes of ontology editing and search for relevant knowledge in the ontology. The possible transitions between the stages of the automaton and which parameters are transmitted at the same time are defined. Such an automaton is the basis for building of GF AFU (Ground Forces of the Armed Forces of Ukraine) tactical units by decision support subsystem commanders.
For modeling the decision support process in the competitive environment the mathematical supply and methods of using the domain ontology in four OODA loop stages (observation, orientation, decision, action) are developed. So for the military sphere on the stage of «Оbservation» the reconnaissance data are analyzed by the subject area ontology in order to determine the strengthes and weaknesses of the enemy. On the «Оrientation» stage the ontological data are used for simulation modeling of the possible course of the fight and for the optimal placement of own forces. In the «Decision» stage the target distribution method based on genetic algorithms, which allows to reduce the computational complexity, and thus significantly speed up the time needed for tactical section commander for decision making. Probability data concerning the destruction of enemy forces with the use of the certain means, which is input data of the developed method, are taken from the ontology based on the analysis of the normative tables. Also, in order to improve the efficiency of possible solutions in the ontology expert knowledge based on the descriptive logic is presented.
Method of using adaptive ontologies in applied subject areas such as military applications taking into consideration the experts definition of ontology individual elements weights, allows us to increase the effectiveness of such Boyd loop stages as «Orientation» and «Decision» during a simulation modeling of the fight flow and target distribution. In some situations, the use of expert rules contained in the ontology allows 20 % probability to increase the own forces intact.
The architecture of decision support subsystem which consists of modules that determine the appropriate stages of the OODA loop is developed. The central component of the decision support subsystem is the subject area ontology. Such ontology of the DF AFU is built. The main elements of ontology is fighting sets, their tactical and technical characteristics, combat statute, normative documents that define the extent of fire damage, and etc. The definition of separate elements of ontology with the use of descriptive logic is made.
The software for each stage of the OODA loop based on the created methods and ontological approach is developed. For the «Observation» stage the mobile application «Military intelligence», which processes the reconnaissance information and transmits it to the command staff is developed. For the realization of the «Orientation» stage the simulation modeling module is developed, the input data of which is the obtained reconnaissance information and the output point of which is the plan of own forces location. For the realization of «Decision» stage the module of the efficient target distribution is developed, the input data of which is the probability of targets attack by certain fire means, and the output is the conformity of fire means to enemy target, which it attacks. For the "Action" stage the mobile application «Adjustment» is developed that corrects shooting. The use of the developed models, methods and software modules gave an opportunity to reduce up to 30 % the time spent by tactical units commanders for planning of military actions.
Key words: decision support system, ontology, knowledge database, Boyd loop, genetic algorithm, mathematical expectation, probability.