Evolving tactics using levels of intelligence in computer-generated forces

Simulated evolution on a computer can provide a means for generating appropriate tactics in real-time combat scenarios. Individual unit or higher-level organizations, such as tanks and platoons, can use evolutionary computation to adapt to the current and projected situations. We briefly review current knowledge in evolutionary algorithms and offer an example of applying these techniques to generate adaptive behavior in a platoon-level engagement of tanks where the mission of one platoon is changed on-the-fly. We also study the effects of increasing the intelligence of one side in a one-on-one tank engagement. The results indicate that measured performance increases with increased intelligence; however, this does not always come at the expense of the opposing side.