Even though combining multi-access techniques (i.e., orthogonal and non-orthogonal multiple access techniques) and multi-band communications (i.e., millimeter-wave and sub-6 GHz communications) is needed to satisfy the stringent requirements of emerging wireless applications, it triggers new resource management challenges. In this paper, we investigate the joint optimization problem of user association, transmit power allocation, and sub-channel assignment in a such network in order to maximize the down-link sum-rate. The problem is formulated mathematically as a mixed-integer non-linear programming problem. Due to the combinatorial and the non-convex characteristics, solving the problem demands further developed complex mathematical tools. To this end, simple but efficient centralized and fully-distributed greedy-based algorithms are proposed, then, genetic algorithms are presented. Simulation results show the efficiency of the proposed solutions and the effectiveness of using heterogeneous multi-access techniques in terms of down-link sum-rate.