| dc.description.abstract | The security companies in Taiwan are mainly in charge of escorting (and delivering) the cash from branches to each major bank and ATM, and also from the companies (i.e. superstore, mall…etc.), which had contract with the security companies, to some specific banks. When the guards from the security company is in charge of the escorting mission of cash-transportation truck, they often faces the risk of robbing, which always causes great property of damage to the clients. In the past 10 years, the robbing incidents in Taiwan have reached up to 35 cases, with an average of 3.5 cars per year being robbed. From the recent cases, we have discovered that, the lack of variability of the route and arriving time of cash-transportation truck is the main factor causing the robberies. Bandit can easily trace, observe, and master the driving tendencies of each truck. Present practical scheduling model is used for artificial scheduling to plan several modes of feasible plans and designs. This mode lacks the systematic analysis which often causes the waste of resources for the resulted scheduling. Besides, the artificial scheduling also lacks the consideration of route variability for the trucks. Therefore, this research constructs a scheduled model for cash-transportation truck, which accounts for the variability of the routes and minimizes the operating costs of security companies as much as possible. Therefore, the research designs parameters of dissimilar space-time, which hoped to help the decision-makers plan the daily route and arriving time as different as possible, and to increase the variability and security of the routes and arriving time for cash-transport truck.
We employ network flow technique to develop the model, which is characterized as NP-hard to formulate the vehicles in the dimension of time and space. We use C++ and CPLEX to solve the problem, but it is difficult to solve in a reasonable time when problem size is huge. Therefore, we develop a heuristic for solving problem effectively. Numerical tests based on real operating data from a security company and reasonable example to evaluate the proposed solution algorithm. The preliminary results could be useful for security companies.
| en_US |