In this study, we tackle the problem of searching for the most favorable pattern of link capacities that makes a power transmission network resilient to cascading failures with limited investment costs. This problem is formulated within a combinatorial multi-objective optimization framework and tackled by evolutionary algorithms. A power flow model (namely, the ORNL-Pserc-Alaska (OPA) model) is embraced to simulate cascading failures in a network and to quantify its resilience. The framework of capacity allocation optimization is originally applied to the 400kV French power transmission network for the purpose of exemplification. The results show that cascade resilient networks tend to have a non-linear capacity-load relation: in particular, heavily loaded components have smaller unoccupied portion of capacity, whereas lightly loaded links present larger unoccupied portion of capacity which is in contrast with the linear capacity-load relation hypothesized in previous works.