Due to the climate changes in recent years, the number and severity of natural disasters have increased significantly, which have caused blackouts in electricity grids. In order to reduce blackouts during these disasters, some solutions have been proposed, among which we can mention the microgrids expansion, taking into account the concept of resilience. Therefore, in this paper, cooperative generation expansion planning model is developed for minimizing the total cost of planning, operation and resilience in a network of microgrids. In the bi-level model developed in this paper, investment decisions are made at a higher level and operational and resilience issues are addressed at the lower level. In order to prove the effectiveness of the developed model, case studies have been conducted in three scenarios including islanding, non-cooperative and cooperative approaches. The results have demonstrated that with the increase of transactions and trading between microgrids, their overall costs have decreased. As a result, the overall costs of microgrids were reduced by 11% and 18% in non-cooperative and cooperative scenarios, respectively; compared to the islanding scenario. Also, due to the importance of environmental issues, the problem has been solved again by considering the objective function of environmental pollutions. In this case, the multi objective optimization problem, including the objective functions of total cost and environmental pollution, has been solved using fuzzy decision-making and multi-objective improved golden ratio optimization algorithm based on the Pareto front.