摘要： One of the advantages of the optical code division multiple access (OCDMA) system networks over other multiple access methods is the ability to provide differentiated classes of quality of service (QoS) for different users. The performance of the OCDMA networks is mainly limited by the multi-access interference (MAI) coming from other users sharing the same spectral resources. The suitable resource allocation in OCDMA networks allows greater capacity and the reliable deployment of the available resources in the network. This work aiming at applying and analyzing four analytical optimization methods applied to energy efficiency (EE) maximization of OCDMA networks. Hence, such optimization methods were applied and their performance-complexity tradeoffs are compared, namely the augmented Lagrangian method (ALM) [18], sequential quadratic programming method (SQP), majoration-minimization (MaMi) approach [23], [24], as well as Dinkelbach’s method (DK). Tests were performed considering practical and realistic OCDMA networks with a wide range of nodes, typically K = {8; 16; 32} nodes. The evaluated optimization methods were able to achieve convergence and perform suitably to solve the optical EE-OCDMA optimization problem. However, in a few specific high loading system configurations, the DK method did not reach full convergence; besides DK processing time was much greater compared to the promising ALM, SQP and MaMi EE-OCDMA optimization approaches.