One of the most important parts when any accident occurs is the issue of control and management before and after the disaster, and if this issue is not taken into account, another secondary catastrophe will occur within the main catastrophe. Therefore, in this study, a random integrated model is proposed. In which two categories of decisions are considered; decisions of the first stage of determining and controlling the inventory of distribution centers and location of distribution centers for pre-disaster and decisions of the second stage for post-disaster, including reviewing the flow of injured, Corpses, homeless and relief items in the network and the allocation of vehicles for this flow. Objectives of this paper include (1) maximizing the probability of successful passage of routes by increasing the reliability of routes, (2) minimizing relief costs pre-disaster and post-disaster, by considering the time window, (3) Minimize maximum uncovered demand for relief goods, for all centers in possible scenarios and situations. Considering the scenario uncertainty situation along with the uncertainty of the route and demand, multi-commodity, several modes of transportation and covering tour are among the innovations of this research. To validate the proposed model in small and medium dimensions, Epsilon restriction method in Gams software environment and for case study (Region 1 of Tehran) in large dimensions has been solved using Invasive Weed Optimization algorithm. The results of the analysis indicate that the Invasive Weed Optimization algorithm will be able to solve the model with the least error compared to the exact solution and less time. Also, as the capacity of distribution centers increases, the cost decreases. and as demand increases, the number of established distribution centers increases, and as the coverage radius increases, the length the tour decreases but number of unoccupied accident hotspots increases and relief costs, including transportation costs, increase