Robust Reward Placement under Uncertainty
Authors: Petros Petsinis, Kaichen Zhang, Andreas Pavlogiannis, Jingbo Zhou, Panagiotis Karras
IJCAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We corroborate our theoretical analysis with an experimental evaluation on synthetic and real data. In this section we evaluate the running time and performance of algorithms on synthetic and real-world data. |
| Researcher Affiliation | Collaboration | Petros Petsinis1 , Kaichen Zhang2 , Andreas Pavlogiannis1 , Jingbo Zhou3 and Panagiotis Karras4,1 1Department of Computer Science, Aarhus University 2Artificial Intelligence Thrust, Hong Kong University of Science and Technology (Guangzhou) 3Business Intelligence Lab, Baidu Research 4Department of Computer Science, University of Copenhagen |
| Pseudocode | Yes | Algorithm 1 Ψ-Saturate Algorithm Input: MMMs Π, steps K, budget L, precision ϵ, param. β. Output: Reward Placement S R of cost at most βL. |
| Open Source Code | Yes | We implemented6 all methods in C++ 17 and ran experiments on a 376GB server with 96 CPUs @2.6GHz. 6https://anonymous.4open.science/r/RRP-F6CA |
| Open Datasets | No | We use two different types of synthetic datasets to represent stochastic networks (i.e., MMMs). In each type, we generate a directed graph and then sample edge weights from a normal distribution to create different settings. We gathered movement records from Baidu Map, covering Xuanwu District in Nanjing from July 2019 to September 2019; The paper does not provide a specific link, DOI, or formal citation for public access to these datasets. |
| Dataset Splits | No | The paper does not explicitly provide details about training, validation, or test dataset splits; it describes synthetic data generation and usage of real-world data without specifying these partitions. |
| Hardware Specification | Yes | We implemented6 all methods in C++ 17 and ran experiments on a 376GB server with 96 CPUs @2.6GHz. |
| Software Dependencies | No | The paper states “We implemented6 all methods in C++ 17” but does not list specific versions for any other software dependencies, libraries, or solvers used in the experiments. |
| Experiment Setup | Yes | We use different problem parameters as Table 1 shows, marking the default value of each parameter in bold. To satisfy the budget constraint for the Ψ-Saturate algorithm, we fix β = 1 as in Corollary 7 and set precision to ϵ = (|Π| 103) 1. We set the budget L as a percentage of the total cost P s S c[s]. |