Current observations (e.g., Colombo et al., 2014) spatially resolve nearby galaxies and indicate that the slope of the giant molecular cloud (GMC) mass function varies between regions within individual galaxies. Magnetohydrodynamics simulations (e.g., Inoue & Inutsuka 2008; Heitsch et al., 2009) have revealed that multiple episodes of compression are essential to form molecular clouds from magnetized warm neutral medium. Inutsuka et al. 2015 proposed that such multiple compressions and the resultant molecular cloud formation are driven by interactions of dense gas on the surface of expanding shells driven by H II regions. Based on this scenario, they formulate a continuity equation for the GMC mass to explore the time evolution of the GMC mass functions, which reproduce the observed variation in the slope of the GMC mass functions. In this seminar, I will present my work that revises the formulation from Inutsuka et al. 2015 by including terms due to cloud-cloud collisions (CCC). Although recent observations suggest that CCC is important for the formation of massive stars and star clusters, my results indicate that CCC does not modify GMC mass function significantly.