Scholarlink

<p>We introduce a general framework for multi-objective decision-making designed to generate ranked lists of candidates that balance competing objectives. The framework is versatile and applicable to a wide range of scenarios where the goal is to select and order a sublist of candidates such that the sequence of their attributes aligns with predefined ideal trends. While broadly applicable, we focus on social choice settings, leveraging the framework to address complex group decision-making tasks that integrate agent preferences (e.g., as expressed through approval ballots) with overarching global constraints. We establish that the problem is generally computationally intractable. To tackle the computational intractability of the general case, we develop heuristic algorithms and evaluate their performance through simulations. To demonstrate the framework’s capabilities, we apply it to two illustrative use cases: democratic playlist editing and democratic job scheduling. In these scenarios, the aim is to generate ranked lists that reflect agent preferences for musical tracks or job tasks, respectively, while adhering to soft constraints on sequencing and attribute transitions over time. By explicitly balancing the competing objectives of preference representation and temporal alignment, the framework demonstrates its ability to handle decision-making challenges that existing models fail to address effectively. Our results showcase its potential not only for social choice but also for other complex decision-making applications, demonstrating its adaptability and utility in solving problems that require harmonizing multiple objectives (A preliminary version of this paper appeared in the 20th European Conference on Multi-Agent Systems (EUMAS 2023), held in Naples&#xa0;<a href="https://eumas23.github.io/home/">https://eumas23.github.io/home/</a>. The current full and revised manuscript extends that work significantly. In particular, the paper now includes democratic scheduling alongside the democratic playlist problem, demonstrating the broader applicability of the model across different scenarios. Additionally, detailed descriptions and discussions of the model have been added, along with further examples and simulation results).</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multiple attribute list aggregation with applications in collaborative playlist editing and collaborative job scheduling

  • 期刊论文

  • 发表日期：

    2026年3月16日

查看全文
  • Eyal Briman,
  • Nimrod Talmon

摘要

We introduce a general framework for multi-objective decision-making designed to generate ranked lists of candidates that balance competing objectives. The framework is versatile and applicable to a wide range of scenarios where the goal is to select and order a sublist of candidates such that the sequence of their attributes aligns with predefined ideal trends. While broadly applicable, we focus on social choice settings, leveraging the framework to address complex group decision-making tasks that integrate agent preferences (e.g., as expressed through approval ballots) with overarching global constraints. We establish that the problem is generally computationally intractable. To tackle the computational intractability of the general case, we develop heuristic algorithms and evaluate their performance through simulations. To demonstrate the framework’s capabilities, we apply it to two illustrative use cases: democratic playlist editing and democratic job scheduling. In these scenarios, the aim is to generate ranked lists that reflect agent preferences for musical tracks or job tasks, respectively, while adhering to soft constraints on sequencing and attribute transitions over time. By explicitly balancing the competing objectives of preference representation and temporal alignment, the framework demonstrates its ability to handle decision-making challenges that existing models fail to address effectively. Our results showcase its potential not only for social choice but also for other complex decision-making applications, demonstrating its adaptability and utility in solving problems that require harmonizing multiple objectives (A preliminary version of this paper appeared in the 20th European Conference on Multi-Agent Systems (EUMAS 2023), held in Naples https://eumas23.github.io/home/. The current full and revised manuscript extends that work significantly. In particular, the paper now includes democratic scheduling alongside the democratic playlist problem, demonstrating the broader applicability of the model across different scenarios. Additionally, detailed descriptions and discussions of the model have been added, along with further examples and simulation results).