Large-scale networked multi-agent systems are becoming complex and highly integrated. To control and optimize such a system would require the use of algorithms that are flexible, scalable, effective, efficient and robust. This dissertation considers the development and application of distributed techniques capable of handling control/optimization problems of these systems. Specifically, problems in the areas of power and energy, and resource allocation are investigated. They are solved as consensus-based optimization problems in situations where communication and collaboration are possible. For systems that cannot be accurately described by a model, the development of model-free distributed control techniques is proposed for system optimization. In this thesis, the design and implementation of model-free distributed algorithms using extremum seeking control is presented for the solution of three different optimization problems. First, the power maximization of an offshore wind farm under fixed and varying free stream wind speeds and directions. Second, the power minimization of a heating, ventilation and air condition system, precisely, a two rooftop air conditioning system with continuously varying capacities (the compressors modulate rather than turn off and on). Finally, the minimization of a resource allocation problem for a class of multi-agent dynamic system subject to some resource constraint. When a model exist for the large-scale system considered, model-based distributed algorithms are favourable. This thesis proposes the development and implementation of a model-based method, a Newton consensus distributed technique for the solution of a multi-agent multi-resource allocation problem. This technique, a second-order method overcomes some of the challenges associated with using first-order methods such as extremum seeking control in the following ways. First, it takes away the need to impose feasible initial conditions. Second, it speeds up convergence and thirdly, reduces the effect of the communication network structure on the performance of the system.