The paper deals with the problem of 1.5 dimensional bin packing. As a data structure carrying information about packaging, a sequence of numbers of rectangles is used, representing the order of their packing. An essential role in obtaining the solution is played by a decoder, which performs the laying of rectangles according to the rules laid down in it. New methods for solving the packing problem are proposed, using mathematical methods in which the principles of natural decision-making mechanisms are laid. Unlike the canonical paradigm ant algorithm to find solutions to the graph G = (X, U) is constructed with a partition on the route of the formation and on the tops within each part, subgraphs whose edges are delayed pheromone. The structure of the solution search graph, the procedure for finding solutions on the graph, the methods of deposition and evaporation of pheromone are described. The time complexity of the algorithm, experimentally obtained, practically coincides with the theoretical studies and for the considered test problems is O(n 2 ). In comparison with existing algorithms, the improvement of results is achieved by 2-3%.