Scheduling Conservation Designs for Maximum Flexibility via Network Cascade Optimization

Authors: Shan Xue, Alan Fern, Daniel Sheldon

JAIR 2015 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate the approach on both real and synthetic conservation data with a standard population spread model. The algorithm is shown to produce near optimal results and is much more scalable than more generic off-the-shelf optimizers. ... 6. Experiments In this section, we first evaluate our primal-dual algorithm by applying it to a real, full-scale conservation problem. Next, to verify the robustness of our approach to other problems, we present results using synthetic conservation data from a problem generator used in several recent studies.
Researcher Affiliation Academia Shan Xue EMAIL Alan Fern EMAIL School of EECS, Oregon State University Corvallis, OR 97331 USA Daniel Sheldon EMAIL School of Computer Science, University of Massachusetts Amherst, MA 01003, USA
Pseudocode Yes Algorithm 1 Primal-Dual Algorithm for SW-DSG. 1: {Inputs: Graph G = (V, E), edge sets E = {E1, . . . , EM}, costs {c1, , c M}, terminals T V} 2: Initialize: uk i = 0, for each k T, i V; wk i,j = 0, for each (i, j) E, k T G = (V, A) with A = lower Bound = 0, solution =
Open Source Code No The paper does not contain any explicit statements about the release of its own source code or links to a repository for the methodology described.
Open Datasets Yes The real map we use is the same dataset as in prior work by Sheldon et al. (2010) on computing upfront conservation designs. ... To evaluate the primal-dual algorithm more thoroughly, we randomly selected 10 synthetic maps generated and used in prior work (Ahmadizadeh et al., 2010).
Dataset Splits No In particular, in our experiments we formed the initial schedules based on a set of 10 cascade scenarios and conducted the pruning step with respect to 40 cascade scenarios.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper mentions using 'CPLEX solver' as a comparison tool, but it does not specify a version number for CPLEX or any other key software dependencies used in their own implementation.
Experiment Setup Yes Our method also requires specifying a strictly decreasing function for defining the surrogate cost function, for which we use f(t) = βt with β = 0.96. ... The results are not very sensitive to the value of β. ... Thus, the remainder of our experiments use 10 cascades for the SAA. ... We compute the reward ˆR(π) by running 20 simulations of the stochastic population spread model. ... Here, we consider problems based on 10 cascades and horizons ranging from 20 to 100 years.