Recent radio observations show that giant molecular cloud (GMC) mass functions noticeably vary across galactic disks (e.g., Colombo et al. 2014). High-resolution magnetohydrodynamics simulations show that multiple episodes of compression are required for creating a molecular cloud in the magnetized interstellar medium (e.g., Inoue et al. 2012). To understand time evolution of GMC mass functions, we formulate the evolution equation for the GMC mass function to reproduce the observed profiles, for which multiple compressions are driven by a network of expanding shells due to H II regions and supernova remnants. We also introduce the cloud-cloud collision (CCC) terms in the evolution equation in contrast to previous work. In this seminar, I would like to present computed time evolutions and the following two suggestions: (1) the GMC mass function slope is governed by the ratio of GMC formation timescale to its dispersal timescale whereas the CCC effect is limited only in the massive end of the profile, (2) almost all of the dispersed gas contributes to the mass growth of pre-existing GMCs in arm regions whereas less than 60 percent contributes in inter-arm regions. Our results suggest that measurement of the GMC mass function slope provides a powerful method to constrain those GMC timescales and the gas resurrecting factor in various environments across galactic disks.