A Framework for Approval-Based Budgeting Methods
Authors: Nimrod Talmon, Piotr Faliszewski2181-2188
AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We define and study a general framework for approval-based budgeting methods and compare certain methods within this framework by their axiomatic and computational properties. Furthermore, we visualize their behavior on certain Euclidean distributions and analyze them experimentally. (...) To compare these methods, we (1) consider their computational complexity; (2) define several axioms, relevant to budgeting methods, and study how well these axioms are satisfied by the methods at hand; and (3) report on three experiments: (...) and (4) an evaluation of certain methods within our framework according to their axiomatic, computational, and visual properties, as well as their ability to deal with local and global items. |
| Researcher Affiliation | Academia | Nimrod Talmon Ben-Gurion University Be er Sheva, Israel talmonn@bgu.ac.il; Piotr Faliszewski AGH University Krakow, Poland faliszew@agh.edu.pl |
| Pseudocode | No | The paper describes algorithmic approaches (Max rules, Greedy rules) in prose and via mathematical formulations but does not include any explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide any specific statement about releasing source code for the described methodology, nor does it include a link to a code repository. |
| Open Datasets | No | The paper describes generating synthetic data for its experiments (e.g., "voters, positioned uniformly on a disc...", "50 cheap items, positioned uniformly..."). It does not use or provide access information for a publicly available, pre-existing dataset. |
| Dataset Splits | No | The paper describes the generation of synthetic data and experimental setups (e.g., distributions of voters and items, budget limits), but it does not specify explicit training, validation, or test dataset splits in terms of percentages or sample counts, as might be found in typical machine learning experiments. The experiments are presented as simulations over generated scenarios. |
| Hardware Specification | No | The paper states that "simulations reported below were performed using the Gurobi ILP solver," but it does not specify any hardware details such as CPU models, GPU types, or memory specifications. |
| Software Dependencies | Yes | Indeed, the simulations reported below were performed using the Gurobi ILP solver (Gurobi Optimization 2018) on such ILP formulations. |
| Experiment Setup | Yes | Experiment 1 and 2 (...) 1. voters, positioned uniformly on a disc of radius 0.3, centered at position (0.5, 0.5); we have 50 such voters for Experiment 1 and 100 for Experiment 2; 2. 50 cheap items, positioned uniformly on a disc of radius 0.2, centered at (0.3, 0.5), and 50 expensive items, positioned uniformly on a disc of radius 0.2, centered at (0.7, 0.5); 3. the cheap items cost 10 each, while the expensive items cost 100 each for Experiment 2, and for Experiment 1, we use a parameter x for the cost of the expensive items (so Table 2 shows histograms for various values of x); 4. the budget limit is 1000 for Experiment 1 and 200 for Experiment 2; 5. for Experiment 1: Each voter approves the 10 items which are the closest to her; for Experiment 2: The approval sets of the voters are generated with respect to a parameter x, as follows: for each cheap item, we identify the 5 voters which are the closest to it, and add the item to their approval sets; for each expensive item, we identify the x voters which are the closest to it, and add the item to their approval sets. |