Tpicola: a Novel Column Based Algorithm for Par- Tial Input Encoding Problems

Recently different design tasks have been formulated as problems of encoding of symbols but restricted to use a given number of binary variables (Partial Encoding Problems). Up to now specific approaches have been developed only for one type of partial encoding problem: the Partial Input Encoding (PIE) problem and have been shown to produce better results than accommodating standard encoding algorithms to that partial problem. Optimization steps, such as the state assignment, which are formulated as Partial Input-Output problems (PIOE) are solved using PIE algorithms or standard input-output algorithms because no specific one has been proposed. This paper presents an original column based approach for the PIE problem that unlike existing ones can be extended to the PIOE. A novel study of unfeasible constraint implementation has been developed too as it is a key factor of the new method. Our preliminary algorithm performs similar to the best previous algorithms while it is portable to the PIOE Problem.