Executive Summary

This interim report reflects the first of a two-phase effort in support of the Defense Modeling and Simulation Office (DMSO) project to review the state of the art in human behavior representation as applied to military simulations. It presents an overview of the needs for models of human behavior, summarizes a methodology for ensuring development of useful models, and describes selected psychological process models that can improve the realism with which human-influenced action is represented.

The need to represent the behavior of individual combatants as well as teams and larger organizations is expanding as a result of increasing use of simulations for training, systems analysis, systems acquisition, and command decision aiding. Both for training and command decision aiding, the behaviors that are important to represent realistically are those that can be observed by the other participants in the simulation, including physical movement and detection and identification of enemy forces. It is important that observable actions be based on realistic decision making and that communications, when they originate with a simulated unit, be interpretable as the result of sensible plans and operations. A team should manifest the range of behaviors required to be consistent with the degree of autonomy it is assigned, including detection and responding to expected and unexpected threats. It should be capable of carrying out actions on the basis of communications typically received from its next higher-echelon commander.

In the panel's view, achieving realism with respect to these observable outcomes requires that the models be based on psychological and sociological theory. For individual combatants, it is important to represent the processes underlying



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 1
--> Executive Summary This interim report reflects the first of a two-phase effort in support of the Defense Modeling and Simulation Office (DMSO) project to review the state of the art in human behavior representation as applied to military simulations. It presents an overview of the needs for models of human behavior, summarizes a methodology for ensuring development of useful models, and describes selected psychological process models that can improve the realism with which human-influenced action is represented. The need to represent the behavior of individual combatants as well as teams and larger organizations is expanding as a result of increasing use of simulations for training, systems analysis, systems acquisition, and command decision aiding. Both for training and command decision aiding, the behaviors that are important to represent realistically are those that can be observed by the other participants in the simulation, including physical movement and detection and identification of enemy forces. It is important that observable actions be based on realistic decision making and that communications, when they originate with a simulated unit, be interpretable as the result of sensible plans and operations. A team should manifest the range of behaviors required to be consistent with the degree of autonomy it is assigned, including detection and responding to expected and unexpected threats. It should be capable of carrying out actions on the basis of communications typically received from its next higher-echelon commander. In the panel's view, achieving realism with respect to these observable outcomes requires that the models be based on psychological and sociological theory. For individual combatants, it is important to represent the processes underlying

OCR for page 1
--> the observable behavior, including sensing and perception, situation awareness, working memory, long-term memory for representing declarative knowledge, decision making, and task management. For larger units, it is important to represent the organizational structure as well as the products of those structures. Added realism also can be achieved by representing a number of individual and organizational behavioral moderators. Moderators at the individual level, such as workload and emotional stress, serve to enhance or degrade performance, as reflected in the speed and accuracy of performance. Moderators at the organizational level, including the average level of training, whether standard operating procedures are followed, the level and detail of those procedures, and the degree of coupling between procedures, all affect performance. It is also important, in the panel's view, to introduce learning and intrinsic variability into human behavior representations. Not only will learning allow human behavior representations to respond realistically to changes in the environment without having to anticipate every possible change ahead of time, but it will also allow the models to more closely reflect intelligent human behavior. Furthermore, capturing different levels of learning will allow the representation of different levels of skill, from novice to expert. Real human behavior is inherently variable and unpredictable, sometimes even irrational. The same person does not always respond identically to identical situations. Although this kind of variability is supported by learning models, it may also be appropriate to introduce controlled randomness into human behavior representations. Trainees who are repeatedly exposed to the same simulations can be expected to learn their idiosyncrasies. The addition of learning and variability will tend to focus them on valid general rules. When undertaking development of human behavior representation, the panel suggests that the Defense Modeling and Simulation Office encourage developers to make use of a systematic methodology. Such a methodology should include the following steps. First, they should employ interdisciplinary teams consisting of sociologists and cognitive psychologists, computer scientists who are knowledgeable in the contemporary literature and modeling techniques, and specialists in the military doctrine and procedures of the domain to be modeled. Second, they should review alternatives and adopt a general architecture that is most likely to be useful for the dominant demands of the specific situation. Third, they should review available unit-level frameworks and support the development of a comprehensive framework for representing the command, control, and communication structure. Fourth, they should review available documentation and seek to understand in depth the domain and its doctrine, procedures, and constraints. They should prepare formal task analyses that describe the activities and tasks as well as the information requirements and the human skill requirements that must be represented in the models. Also, they should prepare unit-level task analyses that describe such factors as resource allocation and communication protocols. Both individual and unit-level analyses should be guided by the computational

OCR for page 1
--> structure that will encode the tasks. Fifth, to prepare estimates of the parameters and variables included in the model that are unconstrained by the domain or procedural requirements, they should use results from the behavioral research literature, procedural model analysis, ad hoc experimentation, and, only as a last resort, expert judgment. Finally, they should systematically test, verify, and validate the behavior and performance of the models at each stage of development. Methods for achieving validation of large, complex models are themselves in need of further research exploration. In the final section of the report the panel identifies several promising substantive areas in which modeling technologies are sufficiently advanced that developers of human behavior representations should investigate and seek to incorporate them in their models. In the area of individual learning, memory, and retrieval, three technologies are discussed—production rules, exemplar, and neural network models; for unit-level learning, the technologies include simulated annealing, neural networks, and genetic programming. In the area of attention and performance, the state of research on low-level attention (governing perception and motor actions) and higher-level planning and decision making are reviewed. In the area of decision making at higher command levels, models derived from game theory, dynamic decision trees, and multiattribute decision theory are presented along with current approaches in operations research, engineering, and organizational science. At the individual combatant level, the discussion focuses on dynamic decision models that allow for individual differences in decision strategy and motivational factors. In the area of situation awareness, there have been fewer attempts to generate computational models, except for some preliminary work on prescriptive models based on belief networks. At the organizational level, most computational models have been multiagent models. They range from symbolic distributed artificial intelligence models to complex adaptive agent models using techniques from genetic algorithms, neural networks, and simulated annealing. In this first phase of its work, the panel concludes that: (1) human behavior representation is essential to successful use of both war games and distributed interactive simulation; (2) current human behavior representations can be improved by incorporating the knowledge of what is already known from the social and behavioral sciences, cognitive science, and human performance modeling; and (3) great additional progress can be expected through the funding of new research and through the application of existing research in areas that the panel is charged to identify both in this interim report and in its final report. In the second phase of the project, the panel will conduct a more thorough review and analysis of the theoretical and applied research of human behavior modeling as it applies to the military context at the individual and larger-unit levels. The panel's final report will recommend a research and development agenda to move the field forward in a systematic and integrated manner. The panel recognizes that its initial articulation of needs in this first phase is broad.

OCR for page 1
--> Drawing on the military background and knowledge of our sponsor, we plan, as part of the second phase, to conduct a more complete analysis of the strengths and weaknesses of current models and assess the feasibility of establishing priorities for human behavior representation in military modeling and simulation activities.