Interactive dynamic influence diagrams (I-DIDs) graphically visualize a sequential decisionproblem for uncertain settings where multiple agents interact not only amongst themselves butalso with the environment that they are in. Algorithms currently available for solving these I-DIDsface the issue of an exponentially growing candidate model space ascribed to the other agents, overtime. One such algorithm identifies and prunes behaviorally equivalent models and replaces themwith a representative thereby reducing the model space. We seek to further reduce the complexityby additionally pruning models that are approximately subjectively equivalent. Toward this, wedefine subjective equivalence in terms of the distribution over the subject agents future actionobservationpaths, and introduce the notion of epsilon-subjective equivalence. We present a new approximationtechnique that uses our new definition of subjective equivalence to reduce the candidatemodel space by pruning models that are epsilon-subjectively equivalent with representative ones.
