Junction tree representations for linear block codes

We introduce a combinatorial representation for linear block codes, called the junction tree representation, which generalizes the notion of code trellis. We first present an algorithm for finding a minimum complexity junction tree. We then show by example that the minimum complexity junction tree can be less complex than the minimal trellis. One implication of this is that one can sometimes devise exact decoding algorithms which have lower complexity than those associated with the minimal trellis.