Saved in:
Bibliographic Details
Main Authors: Xu, Peipei, Ma, SiYuan, Liu, Yaohua, Wu, Yu, Liu, Guanliang, Zhang, Yang, Liu, Yong
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.12106
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918497117798400
author Xu, Peipei
Ma, SiYuan
Liu, Yaohua
Wu, Yu
Liu, Guanliang
Zhang, Yang
Liu, Yong
author_facet Xu, Peipei
Ma, SiYuan
Liu, Yaohua
Wu, Yu
Liu, Guanliang
Zhang, Yang
Liu, Yong
contents Generating feasible Pareto fronts for constrained bi-objective continuous optimization is central to multi-criteria decision-making. Existing methods usually rely on iterative scalarization, evolutionary search, or problem-specific solvers, requiring repeated optimization for each instance. We introduce DIPS, an end-to-end framework that fine-tunes large language models as amortized Pareto-front generators for constrained bi-objective convex optimization. Given a textual problem description, DIPS directly outputs an ordered set of feasible continuous decision vectors approximating the Pareto front. To make continuous optimization compatible with autoregressive language modeling, DIPS combines a compact discretization scheme, Numerically Grounded Token Initialization for new numerical tokens, and Three-Phase Curriculum Optimization, which progressively aligns structural validity, feasibility, and Pareto-front quality. Across five families of constrained bi-objective convex problems, a fine-tuned 7B-parameter model achieves normalized hypervolume ratios of 95.29% to 98.18% relative to reference fronts. With vLLM-accelerated inference, DIPS solves one instance in as little as 0.16 seconds and outperforms general-purpose and reasoning LLM baselines under the evaluated setting. These results suggest that LLMs can serve as effective amortized generators for continuous Pareto-front approximation.
format Preprint
id arxiv_https___arxiv_org_abs_2605_12106
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Large Language Models as Amortized Pareto-Front Generators for Constrained Bi-Objective Convex Optimization
Xu, Peipei
Ma, SiYuan
Liu, Yaohua
Wu, Yu
Liu, Guanliang
Zhang, Yang
Liu, Yong
Artificial Intelligence
Generating feasible Pareto fronts for constrained bi-objective continuous optimization is central to multi-criteria decision-making. Existing methods usually rely on iterative scalarization, evolutionary search, or problem-specific solvers, requiring repeated optimization for each instance. We introduce DIPS, an end-to-end framework that fine-tunes large language models as amortized Pareto-front generators for constrained bi-objective convex optimization. Given a textual problem description, DIPS directly outputs an ordered set of feasible continuous decision vectors approximating the Pareto front. To make continuous optimization compatible with autoregressive language modeling, DIPS combines a compact discretization scheme, Numerically Grounded Token Initialization for new numerical tokens, and Three-Phase Curriculum Optimization, which progressively aligns structural validity, feasibility, and Pareto-front quality. Across five families of constrained bi-objective convex problems, a fine-tuned 7B-parameter model achieves normalized hypervolume ratios of 95.29% to 98.18% relative to reference fronts. With vLLM-accelerated inference, DIPS solves one instance in as little as 0.16 seconds and outperforms general-purpose and reasoning LLM baselines under the evaluated setting. These results suggest that LLMs can serve as effective amortized generators for continuous Pareto-front approximation.
title Large Language Models as Amortized Pareto-Front Generators for Constrained Bi-Objective Convex Optimization
topic Artificial Intelligence
url https://arxiv.org/abs/2605.12106