Cookies

We use cookies to ensure that we give you the best experience on our website. You can change your cookie settings at any time. Otherwise, we'll assume you're OK to continue.

Durham University

Research & business

Research lectures, seminars and events

The events listed in this area are research seminars, workshops and lectures hosted by Durham University departments and research institutes. If you are not a member of the University, but  wish to enquire about attending one of the events please contact the organiser or host department.


 

December 2020
SunMonTueWedThuFriSat
November 2020 January 2021
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31

Events for 1 December 2020

Riccardo Mogre: Dynamic Project Expediting: A Stochastic Shortest-Path Approach

1:00pm, Online

A decision maker is in charge of a project whose progress is random because of disruptions or productivity problems. In each time period, the manager reviews the progress and decides whether to expedite each activity. Her problem is to identify expediting policies that minimize her expected cost, given by the sum of the cost of expediting and a cost linear in the project completion time. We propose a infinite-horizon Markov decision process to identify optimal expediting policies for the execution of projects. We show that our problem belongs to a class of stochastic shortest-path problems which have some special ordering properties, which we call “forward-only stochastic shortest-path problems.” The enumeration of the feasible states for our problem is very difficult, primarily because of the precedence constraints in the network. For this reason, we devise algorithms to identify all the feasible states of the problem. The complexity of this problem renders impractical the use of existing algorithms employed to solve stochastic shortest-path problems. For this reason, we devise an exact, computationally-efficient algorithm to solve forward-only stochastic shortest-path problems. We complement our analytical results with a computational study that shows the computation times for various randomly generated networks.

Contact algorithms.complexity@dur.ac.uk for more information about this event.