A robust-stochastic optimisation approach for a logistics planning problem

Planning for training events is a challenging task because coordination is necessary. Training sites need to be identified and specific equipment should be at each training site during the events. Often, key decisions such as when and where to host the training event, and how much equipment to send to each site need to be defined much earlier than the starting date of the events and cannot be easily modified later.

In real-life applications experienced by Singapore’s Defence Science and Technology Agency (DSTA), the starting time of each event may often deviate from what was scheduled, and the amount of equipment present at each training site when the event starts is also uncertain. These uncertainties can disrupt the smooth progress of the training event. Hence, it is crucial for an equipment redistribution plan to be devised between the training sites such that equipment requirements of training events are fulfilled as much as possible, even in the presence of uncertain start times and equipment availability.

This Technical Note proposes optimisation models including uncertainty (e.g., robust, stochastic) to help address such challenges. Results show that solutions are more realistic than those obtained with deterministic approaches where uncertainty is not considered.

For further details, the Technical Note Download'A Robust-Stochastic Optimisation Approach for a Logistic Planning Problem' (PDF, 1.6 MB)vertical_align_bottom, is available online.