چكيده به لاتين
The disaster response network design (prevention and preparation phase) has great importance due to the possibility of further damage and the destruction of transport infrastructure due to natural disasters such as earthquakes. The present dissertation has presented a method in the form of an integrated multi-objective optimization model for the emergency response routes network design and performing emergency and effective logistics operations after an earthquake. In the present dissertation, general optimization models in both stochastic and non-stochastic programming problems are presented by introducing and expanding performance indexes (including connection, vulnerability, length, link and route importances). Due to the fact that the above issues are multi-objective in nature, so in order to convert these functions into optimal single-objective models and to prunne the soloutions by increasing the applicability and reducing the role of decision makers, accurate and multi-objective approaches are used. In the proposed approaches fot multi-objective transformation functions, methods such as total weighted method, bounded objective function method, combined total weighted method and lexicographical method were used. A exact branch and cutting method was used to solve the transformed problems. To prunning the solutions, a trade-off index was used for emergency response paths and a lexicographic-based approach, including a combined solution method (relaxation induced neighborhood search, local branching, and constraint planning). These results were presented for the Sioux-Falls transportation network and the large-scale transportation network in Tehran. The results also show the importance of network vulnerability performance criteria for identifying emergency response routes. In the present thesis, the effective parameters in determining the emergency road network have been introduced for the first time in the optimization model and have been presented as an integrated multi-objective model. The results show the efficiency of the proposed models and their applicability to real-scale problems.
