Computing Nash Equilibria in Generalized Interdependent Security Games

Authors: Hau Chan, Luis E. Ortiz

NeurIPS 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we experimentally examine and discuss the practical impact that the additional protection from transfer risk allowed in generalized IDS games has on MSNE by solving several randomly-generated instances of SC+SS-type games with graph structures taken from several real-world datasets.
Researcher Affiliation Academia Hau Chan Luis E. Ortiz Department of Computer Science, Stony Brook University {hauchan,leortiz}@cs.stonybrook.edu
Pseudocode Yes Appendixes A.1 and B of the supplementary material contain our versions of the lemmas and detailed pseudocode for the algorithm, respectively.
Open Source Code No The paper does not provide an unambiguous statement or link indicating that the source code for the described methodology is publicly available.
Open Datasets Yes The underlying structures of the instances use network graphs from publicly-available, real-world datasets [6, 16 20].
Dataset Splits No The paper describes how the experimental instances are generated and how a heuristic is run on them, but it does not specify any training, validation, or test dataset splits or cross-validation setups.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory, cloud instances) used for running the experiments.
Software Dependencies No The paper mentions the use of a 'simple gradient-dynamics heuristic based on regret minimization' but does not specify any software names with version numbers for replication.
Experiment Setup Yes On each instance, we initialize the players mixed strategies uniformly at random and run a simple gradient-dynamics heuristic based on regret minimization [21 23] until we reach an (ϵ) NE. In short, we update the strategies of all non-ϵ-best-responding players i at each round t according to x(t+1) i x(t) i 10 (Mi(1, x(t) Pa(i)) Mi(0, x(t) Pa(i))). Note that for ϵ-NE to be well-defined, all Mis values are normalized. To generate each instance we generate (1) Ci/Li where Ci = 103 (1+random(0, 1)) and Li = 104 (or Li = 104/3) to obtain a low (high) cost-to-loss ratio and αi values as specified in the experiments; (2) pi such that sc i or ss i is [0, 1]; and (3) qji s consistent with probabilistic constraints relative to the other parameters (i.e. pi +P j P a(i) qji 1).