Robust Stackelberg Equilibria in Extensive-Form Games and Extension to Limited Lookahead

Authors: Christian Kroer, Gabriele Farina, Tuomas Sandholm

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

Reproducibility Variable Result LLM Response
Research Type Experimental Using our MIPs presented in the previous section we investigated the scalability and qualitative properties of RSSS solutions. We experimented with three kinds of EFG: Kuhn poker (Kuhn) (Kuhn 1950), a 2-card poker variant (2-card), and a parameterized security-inspired search game (Search). ... All experiments were conducted using Gurobi 7.5.1 to solve MIPs, on a cluster with 8 Intel Xeon E5607 2.2Ghz cores and 47 GB RAM per experiment. In the first set of experiments we investigate the impact on runtime caused by uncertainty intervals in each of the four games... The results are shown in Table 1.
Researcher Affiliation Academia Christian Kroer, Gabriele Farina, Tuomas Sandholm Computer Science Department, Carnegie Mellon University {ckroer,gfarina,sandholm}@cs.cmu.edu
Pseudocode No The paper provides mathematical programs (MIPs) but does not present them as structured pseudocode or algorithm blocks with explicit labels like "Algorithm" or "Pseudocode".
Open Source Code No The paper does not provide any statement or link indicating that the source code for the described methodology is publicly available.
Open Datasets No The paper defines the game structures (Kuhn poker, 2-card poker, Search game) but does not provide concrete access information (links, DOIs, or formal citations) for a pre-existing publicly available dataset used for training or evaluation.
Dataset Splits No The paper does not describe dataset splits (training, validation, test) as it defines game structures rather than using pre-existing datasets that would require splitting.
Hardware Specification Yes All experiments were conducted using Gurobi 7.5.1 to solve MIPs, on a cluster with 8 Intel Xeon E5607 2.2Ghz cores and 47 GB RAM per experiment.
Software Dependencies Yes All experiments were conducted using Gurobi 7.5.1 to solve MIPs
Experiment Setup Yes In the first set of experiments we investigate the impact on runtime caused by uncertainty intervals in each of the four games, without considering limited lookahead. We compare the MIP by Bosansky and Cermak (2015) (B&C) for the full-certainty setting to our robust MIP ((17)-(29)) with an uncertainty interval of diameter d at each node in the game, for 6 different values of d. ... We instantiate the MIP with several different uncertainty-interval widths (given in the leftmost column in Table 2). For each leader strategy, we then conduct a grid search over triplets of numbers in { 0.1, 0.5, 1, 2, 3}3, where the three numbers correspond to a change in utility being added to each of the three rightmost payoff nodes in Figure 2.