Codes can be represented by edge-labeled directed graphs called trellises, which are used in decoding with the Viterbi algorithm. We will first examine the well-known product construction for trellises and present an algorithm for recovering the factors of a given trellis. To maximize efficiency, trellises that are minimal in a certain sense are desired. It was shown by Koetter and Vardy that one can produce all minimal tail-biting trellises for a code by looking at a special set of generators for a code. These generators along with a set of spans comprise what is called a characteristic pair, and we will discuss how to determine the number of these pairs for a given code. Finally, we will look at trellis dualization, in which a trellis for a code is used to produce a trellis representing the dual code. The first method we discuss comes naturally with the known BCJR construction. The second, introduced by Forney, is a very general procedure that works for many different types of graphs and is based on dualizing the edge set in a natural way. We call this construction the local dual, and we show the necessary conditions needed for these two different procedures to result in the same dual trellis.

Two standard graph representations for linear codes are the Tanner graph and the tailbiting trellis. Such graph representations allow the decoding problem for a code to be phrased as a computational problem on the corresponding graph and yield graph theoretic criteria for good codes. When a Tanner graph for a code is used for communication across a binary erasure channel (BEC) and decoding is performed using the standard iterative decoding algorithm, the maximum number of correctable erasures is determined by the stopping distance of the Tanner graph. Hence the computational problem of determining the stopping distance of a Tanner graph is of interest. In this thesis it is shown that computing stopping distance of a Tanner graph is NP hard. It is also shown that there can be no (1 + є ) approximation algorithm for the problem for any є > 0 unless P = NP and that approximation ratio of 2(log n)1- є for any є > 0 is impossible unless NPCDTIME(npoly(log n)). One way to construct Tanner graphs of large stopping distance is to ensure that the graph has large girth. It is known that stopping distance increases exponentially with the girth of the Tanner graph. A new elementary combinatorial construction algorithm for an almost regular LDPC code family with provable Ώ(log n) girth and O(n2) construction complexity is presented. The bound on the girth is close within a factor of two to the best known upper bound on girth. The problem of linear time exact maximum likelihood decoding of tailbiting trellis has remained open for several years. An O(n) complexity approximate maximum likelihood decoding algorithm for tail-biting trellises is presented and analyzed. Experiments indicate that the algorithm performs close to the ideal maximum likelihood decoder.