| Summary |
Tramp ships, like taxis, follow the cargoes which can be obtained at the best rates, and
have no fixed schedules or published ports of call. In contrast, tramp shipping
companies are often faced with high operational risk of uncertain demands and bunker
prices. Hence, it is critical for a tramp shipping company to manage their fleet. The goal
of this study is to help tramp shipping companies dynamically manage their ship fleets
by taking into account ship chartering strategies, uncertain demands and bunker prices.
The research considers the following essential issues: (1) fleet size problem; (2) vessel
routing and scheduling problem with variable speed; (3) ship charter scheme (when and
where to charter ships in and charter ships out?); (4) refueling decision (when and
where to refuel, and how much?); Note that ship fuel consumption is a function of
speed. This research first mathematically models the problem of dynamic tramp ship
fleet management with uncertain demands and bunker prices. However, it is very
difficult and intractable to compute the exact solution by dynamic programming
because of the curse of dimensionality. Hence, this research applies approximate
dynamic programming to tackle this dynamic stochastic problem. This research set the
four dimensional piecewise-linear value function and comes up with a way to setting up
initial values of the value function. And the results show the speed of convergence of
our way is faster than the original way, which just gives the initial value as 0. The last,
we compare our method with benchmark data, and do the experiment by using the
practical data and which comes from a tramp shipping company in Taiwan. |
