IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS, VOL. SMC-12, NO. 4, JULY/AUGUST 1982 [11] A. deGroot, "Perception and memory versus thought: Some old ideas and recent findings," in B. Kleinmuntz, Ed., Problem Solving. New York: Wiley, 1966. [12] David K. Dirlan, "Most efficient chunk sizes," Cognitive Psychol., vol. 3, 1972. [13] J. J. Donovan, "Database system approach to management decision support," ACM Trans. Decision Syst., vol. 1, no. 4, Dec. 1976. [14] P. W. Frey, Ed., Chess Skill in Man and Machine. New York: Springer-Verlag, 1977. [15] P. W. Frey and P. Ademan, "Recall memory for visually presented chess positions," Memory and Cognition, vol. 4, no. 5, 1976. [16] E. F. Lindquist, Design and Analysis of Experiments in Psychology and Education. Boston: Houghton Mifflin, pp. 266-273, 1953. [17] M. Newborn, Computer Chess. New York: Academic, 1975. [18] A. Newell and H. A. Simon, Human Problem Solving. Englewood Cliffs, NJ: Prentice-Hall, 1972. [19] J. S. Reitman, "Skilled perception in GO: Deducing memory struc- tures from inter-response times," Cognitive Psychol., vol. 8, no. 3, pp. 336-356, July 1976. [20] F. Restle and E. Brown, "Organization of serial pattern learning," in G. H. Bower, Ed., The Psychology of Learning and Motivation. vol. 4, New York: Academic, 1970. [21] A. L. Samuel, "Some studies in machine learning using the game of checkers II, Recent progress," IBM J. Res. Develop., vol. 11, no. 6, 1967. [22] V. Slamecka, H. Camp, A. N. Badre, and D. Hall, "MARIS: A knowledge system for internal medicine," J. Information Manage- ment and Processing, vol. 13, pp. 273-276, 1977. [23] A. L. Zobrist and F. R. Carlson, Jr., "An advice-taking chess computer," Sci. Amer., vol. 228, pp. 92-105, 1973. A Trainable On-Line Model of the Human Operator in Information Acquisition Tasks AZAD M. MADNI, MEMBER, IEEE, MICHAEL G. SAMET, AND AMOS FREEDY, MEMBER, IEEE Abstract-A trainable model of the human operator in information acquisition tasks is described. The purpose of this model, called the adaptive information selector (AIS), is to select and present textual mes- sages automatically to users in computer-based tactical systems. AIS algorithms, which control and present messages in order of priority in real time, are based on multiattribute characterization of tactical messages. The AIS employs an adaptive pattern recognition model to "learn" user prefer- ence structure incrementally during actual task performance. Across each of the command situations, the priority of messages is determined by the AIS in accord with the information selection behavior exhibited by the user in the model-training mode. The AIS was implemented and tested in a simulated environment. The results demonstrated the model's capability to 1) converge on distinctive information processing strategies exhibited by different operators; and 2) effectively present messages in their order of priority for each operator. The AIS is potentially useful in performing information distribution functions in command and control systems and in aiding the performance of personalized searches of large data bases. I. INTRODUCTION T ECHNOLOGICAL advances have led to rapid in- creases in the amount and transfer rate of military information. Today information must be processed more efficiently and more effectively for commanders to make tactical decisions responsive to the rapidly changing succession of events. To meet this need, new computer- Manuscript received November 1, 1981; revised March 5, 1982. This work was supported by Defense Advanced Project Agency Cybernetics Technology Office under Contract MDA903-78-C-0127. The authors are with Perceptronics, Inc., 6271 Variel Avenue, Wood- land Hills, CA 91364. based systems for command, control, and communications (C3) are being continually developed, implemented, and evaluated. These systems are intended primarily to aid in the collection, processing, and utilization of different types and amounts of military data. The overall process is cyclic -as information is being used, other information is being processed, and new information is being sought and col- lected. In this environment the dynamics of information flow is of critical importance and must be constantly monitored and directed. The consensus on current computer-based military sys- tems for C3 operations is that the message traffic within the network has increased to such an extent as to over- whelm a commander and his staff. New techniques to control information flow are required to best match system capability with human operator characteristics. Review of previous research has suggested that a significant step in this direction would be to individualize and automate information selection [1]. The goal is to allow each user to obtain information consistently that is both relevant and timely with regard to his individual processing characteris- tics and immediate decisionmaking needs. Considering the large number of users in a typical C3 system, the effect of individualized message handling systems on total system performance would be to both increase throughput and improve decisionmaking performance. Thus the purpose of the research undertaken here was to demonstrate means by which computer-based models of 0018-9472/82/0700-0504$00.75 (©1982 IEEE 504