This work consists on the application of cutting plane techniques (accelerated euclidean algorithm and disjunctive cuts) for solving pure 0-1 integer problems and their specializations for the set partitioning problem, when combined to penalty techniques for fixing variables. A study on penalty techniques, which allows the fixation of variables to integer values, is also developed. These penalties are directly derived from the optimal tableau nof the linear relaxation of the integer problem. The variables fixed due to penalties are eliminated and the problem is reformulated, having its initial dimensions reduced. Some improvements on the evaluation of penalties are suggested, taking into account the special structure of the set partitioning problem. Finally, a new approach to the solution of set partitioning problems is proposed: a cutting plane algorithm which uses penalty techniques, in order to accelerate the convergence of pure cutting plane methods and overcome the problems arising from their use. Computational results are shown, allowing to compare the performance of (i) the accelerated euclidean algorithm, (ii) the disjunctive cut algorithm and (iii) the last one combined with penalty techniques. For the latter, the results obained by the use of generic penalties for 0-1 integer programs are compared with those obtained by the use of the improved penalties for ser partitioning problems. Taking into account set partitioninng problems and the improvements proposed for the evaluation of penalties, it is shown that very often it is possible to fix more variables to integer values and even to solve directly the original 0-1 problem. For some cases, by applying the cutting plane algorithm together with penalties, it is possible to accelerate the convergence and overcome dual degeneracy and round-off errors arising from the use of pure cutting plane algorithms.