In this paper we address the problem of joint transmitter and receiver optimization in a wireless multiuser MIMO-OFDM environment. We consider a centralized design strategy assuming that Channel State Information (CSI) is available. The optimization goal is the minimization of the total transmitted power subject to different constraints, such as the Quality of Service (QoS) for each user in terms of the mean Bit Error Rate (BER) or the maximum power for each transmitter. In this design, the physical resources are used in an optimum way and the Multiple Access Interference (MAI) is minimized. The proposed technique is a variant of the Simulated Annealing (SA) algorithm, which is able to find the global optimum. Besides, it can handle all kind of constraints by generalizing the definition of the cost function. Finally, we present simulation results and compare this technique with a Lagrange-gradient based solution.