In this paper we consider problems of optimization and selection of development systems (technologies) of oil/gas fields, consisting of some disjoint oil/gas pools (in terms of hydrodynamics), tied by resource constraints or general oil/gas production plan. In order to solve these problems, formulated as MILP models, we have developed approximate algorithm using Lagrangian relaxation. Initially we consider the problem for oil fields and then for gas fields. Contrary to another models and techniques, used for solving the similar problems, our models and algorithms allow us to coordinate allocation of production volumes and reserves among the pools with selection of optimal development system, as well as optimization of technological parameters for each pool. We have also examined the perspective approaches, using both multilevel decomposition of oil reservoirs, and hierarchical splitting, and parallel computing on supercomputer for developing effective problem-solving procedures.