چكيده به لاتين
In supply chain management for valuables, secret or hazardous goods, it is not enough to just pay attention to economic goals, and goods' distributors should also consider security goals in order to benefit from the smart and agile chain. For example, in banking industry, if the decisions made regarding the network design and the manner of cash and securities distribution in the physical network are not made correctly, despite the increase in time and cost of transporting cash in the system, it leads to delays in the distribution or endangers the network. It is obvious that the existence of delays and risks in the network or the lack of adequate provision of cash for banks customers will lead to their dissatisfaction and therefore to a decrease in bank customers. For this reason, banks have to use the necessary techniques to properly manage the cash logistics, so that by distributing the proper quality and quantity of cash and securities in the distribution network, in spite of providing suitable customer service, retain current customers or even increase their number. In this regard, this dissertation illustrates, models and solves the issue of secure distribution network design, where at the strategic level pays attention to location/restructuring of network' facilities and at the operational level optimizes the transportation of valuables, secret or hazardous goods in order to achieve the necessary intelligence in economic and security purposes.
In this regard, while explaining the importance of the issues raised, we first describe a new spatial decision support methodology to restructure the branches' network at the strategic level. Using the results, an integrated location routing to locate the treasury and route vehicles is developed taking security criteria into account. This model is further extended by modeling demand fluctuations and traffic conditions in the urban environment, as the two real-world features, using a robust optimization approach and time-dependency of travel speeds, respectively. Due to the NP-hard feature of the routing model resulting from the development of the basic "vehicle routing problem" with the addition of new constraints, objectives and assumptions, two meta-heuristic methods have been developed to solve large-size data. Finally, by performing several computational experiments to validate, evaluate and analyze the models, algorithms and methods developed on three series of artificial data, benchmark data in the literature and real case of branches' network of Bank Melli. The results show the extensive efficiency regarding these three cases. Regarding the efficiency of the developed models on the real case, after restructuring the branches' network in line with the bank's objectives and considering the relevant constraints (the number of branches and strategic considerations that did not exist before), only using the implementation of our LRP model based on the security criteria viewpoint, the improvements of 6.26% and 38.15% are achieved in the objective and risk functions, and from a strategic point of view, in 67% of the data, stability in location decisions is reported. Also, by modeling time-dependent travel speeds, we see an average of 9.31% improvement over constant speed conditions. In addition, protecting the time-dependent model against the worst-case scenario under a maximum uncertainty level of 0.2 results in an average cost increase of 7.45% over the deterministic model.
