IET is a key contributor on DARPA's effort to build enhancements for Cougaar (a distributed agent architecture for military logistic planning) that address the Ultra*Log system survivability by supporting improved robustness using subjective user tradeoffs grounded in probabilistic assessment of logistics planning alternatives. IET's goal was to build a Java plug-in package for the Cougaar library that implements utility-based logistics plan execution monitoring for hierarchical plan fault detection, isolation, and prediction. We have moved towards that goal with enhancements that fall in two related infrastructure areas: Multiple worlds (MW) and Probabilistic Inference (PI). These enhancements provide Ultra*Log/Cougaar users with the ability to develop multiple operation plans (OPlans) simultaneously in different worlds, with each world supporting "what-if" processing, thereby providing a material basis to trade off planning alternatives.