Automated Negotiation with Gaussian Process-based Utility Models
Authors: Haralambie Leahu, Michael Kaisers, Tim Baarslag
IJCAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our numerical experiments are summarized in Table 2, illustrating the dependence, w.r.t. the query cost-factor γ, of the (expected) discounted utility Util and the number of queries Que , for both the myopic and random elicitation approaches, based on averages over 100 (myopic), respectively 400 (random), simulations of the negotiation process. We note that the myopic approach clearly outperforms the random one in terms of discounted expected utility, scoring higher for any cost-factor γ. |
| Researcher Affiliation | Academia | Haralambie Leahu , Michael Kaisers and Tim Baarslag Centrum Wiskunde & Informatica, Amsterdam, The Netherlands {H. Leahu, M. Kaisers, T. Baarslag}@cwi.nl |
| Pseudocode | Yes | Algorithm 1 decides between elicitation and negotiation Algorithm 2 generates a myopic elicitation cycle |
| Open Source Code | No | The paper does not provide any explicit statement about making the source code available, nor does it include a link to a code repository. |
| Open Datasets | No | For our experiments, we choose a negotiation space consisting of 10 items/options, labeled 0, 1, . . . , 9 and use the ordinal utility functions φ (for the user) and ψ (for the opponent) displayed in Table 1. The paper does not provide concrete access information (link, DOI, repository) for this dataset. |
| Dataset Splits | No | The paper describes its experimental setup as simulations using predefined utility functions and compares two elicitation strategies. It does not specify traditional training, validation, or test dataset splits (e.g., percentages or sample counts). |
| Hardware Specification | No | The paper does not mention any specific hardware specifications (e.g., CPU, GPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper does not specify any software dependencies with version numbers. |
| Experiment Setup | Yes | For our experiments, we choose a negotiation space consisting of 10 items/options, labeled 0, 1, . . . , 9 and use the ordinal utility functions φ (for the user) and ψ (for the opponent) displayed in Table 1. ... we assume a distance-based correlation structure k(x, z) = κ(|x z|), with κ(0) = 0.5, κ(1) = 0.3, κ(2) = 0.2, κ(3) = 0.1 and κ = 0 otherwise;... we assume that all queries are equally costly, with each query inducing a discount γ (0, 1] on the final reward |