Šio darbo tikslas – sukurti trimačių dėžių su papildomu orientacijos erdvėje apribojimu pakavimo optimizavimo algoritmą, o realizavus jį programinėmis priemonėmis – ištirti jo efektyvumą su atsitiktinai sugeneruotų kraunamų dėžių rinkiniais ir prie skirtingų algoritmo veikimo parametrų. Taip pat sukurti ir pakavimo optimizavimo sprendinio vizualizavimo trimatėje erdvėje programinę įrangą. / Presented work covers one of the most complex areas of combinatorial optimization – three dimensional bin packing problem. Solution methods of this problem are applied in the real world from logistics, packing optimization to VLSI circuit and automobile engineering. Several heuristic packing algorithms suggested by other authors are analyzed. Approach based on tree-search and wall building strategy is chosen to create a 3D packing optimization algorithm. A bin orientation in space restriction is added to classical 3D bin packing problem to make it more complex and more suited for real world applications. A prototype of created algorithm is created and tested with randomly generated data collections. Each data sample is processed with and without bin orientation in space restriction. Influence of restriction and maximal tree width on packing efficiency and computational time is statistically analyzed. Visualization tool based on Microsoft Direct X technology is created to view results of packing optimization.