Competitors in the marketplace or combatants on the battlefield face very similar challenges: their resources, be they money or weapons, are gradually attrited in the mutual effort to push each other out of the field and dominate it. Even if the participants are deterministic in their decision-making, executing their decisions has random aspects, when the same, generally successful actions occasionally fail for no obvious reasons. The application of system and control theories to improve the planning as well as the execution of such processes requires models, which allow planners and managers to reliably predict the expected outcomes of various alternatives over a long horizon into the future. In this article, exact probabilistic models for several classes of battle scenarios are developed from the first principles, which accurately characterize the battle dynamics for arbitrarily long horizons. It is shown how the models are used for model predictive control of the battle dynamics.
[1]
T. Başar,et al.
Dynamic Noncooperative Game Theory, 2nd Edition
,
1998
.
[2]
Joshua M. Epstein,et al.
Nonlinear Dynamics, Mathematical Biology, And Social Science
,
1997
.
[3]
Edward J. Sondik,et al.
The optimal control of par-tially observable Markov processes
,
1971
.
[4]
Jan Jelinek.
Predictive models of battle dynamics
,
2001,
SPIE Defense + Commercial Sensing.
[5]
Gordon M. Clark,et al.
The combat analysis model
,
1969
.
[6]
J. Filar,et al.
Competitive Markov Decision Processes
,
1996
.
[7]
C. J. Ancker,et al.
A proposed foundation for a theory of combat
,
1995
.
[8]
Richard Hillestad,et al.
Cutting some trees to see the forest: On aggregation and disaggregation in combat models
,
1995
.