Complexity of Scheduling Charging in the Smart Grid

Authors: Mathijs de Weerdt, Michael Albert, Vincent Conitzer, Koos van der Linden

IJCAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental An experimental study establishes up to what parameter values the dynamic programs can determine optimal solutions in a couple of minutes. and In the following experimental analysis we aim to establish when runtimes are sufficiently large that there is significant value in researching faster (but possibly non-optimal) algorithms instead of straightforwardly using the proposed dynamic programs.
Researcher Affiliation Academia Mathijs de Weerdt1, Michael Albert2, Vincent Conitzer2, and Koos van der Linden1 1 Delft University of Technology, Delft, The Netherlands 2 Duke University, Durham, North Carolina, USA
Pseudocode Yes DP1 : OPT(m1, m2, . . . , m|T |, i) = ( 0 if i = 0 max OPT(m1, m2, . . . , m|T |, i 1), o otherwise max a1,.,a|T | OPT(m1 a1, . . . , m|T | a|T |, i 1) + vi (a) . and 1. Sort all charging task triples on deadline (increasing, with arbitrary tie-breaking). 2. Let M1, M2, . . . , Mn be the cumulative supply at the deadlines of tasks 1, 2, . . . , n that is, Mi = Pdi t=1 mt and let M0 = 0. 3. Run a DP based on the following recursion (where m denotes the remaining cumulative supply available for the first i tasks): OPT(m, i) = ...
Open Source Code No No explicit statement or link providing concrete access to the source code for the methodology described in the paper was found.
Open Datasets Yes to model this, we use load measurements of a household with 15 minute resolution from the Pecan Street dataset [Pecan Street Inc., 2017].
Dataset Splits No The paper describes how problem instances are generated randomly with certain parameters, but it does not specify explicit train/validation/test dataset splits or cross-validation setup for these instances or the Pecan Street dataset.
Hardware Specification Yes on a standard desktop computer (in our case containing an i7-6700 3.4Ghz quad core processor with 32GB of memory)
Software Dependencies No The paper states, 'We implemented both DP1 and DP2 in python using a single threaded tabulation approach,' but it does not specify the Python version or any other software dependencies with version numbers.
Experiment Setup Yes We start with relatively small problem instances, using the following default settings (unless stated otherwise): there are 4 agents, the horizon is 6 time steps, supply at each time step is scaled into the range of 3 times the maximum charging speed, and the maximum demand per vehicle is equivalent to charging 3 time steps at full speed. The number of deadlines is either 1 (single) or 3 (multiple deadlines).