Research Report on Automatic Synthesis of Local Search Neighborhood Operators
Constraint Programming (CP) and Local Search (LS) are different paradigms for dealing with combinatorial search and optimization problems. Their complementary features motivated researchers to create hybrid CP/LS solutions, maintaining both the modeling capabilities of CP and the computational advantages of the heuristic-based LS approach. Research presented in this report is focused on developing a novel method to infer an efficient LS neighborhood operator based on the problem structure, as modeled in the CP paradigm. We consider a limited formal language that we call a Neighborhood Definition Language, used to specify the neighborhood operators in a fine-grained and declarative manner. Together with Logic Programming runtime called Noodle, it allows to automatically synthesize complex operators using a Grammar Evolution algorithm.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
HBCP Corpus: A New Resource for the Analysis of Behavioural Change Intervention Reports
Due to the fast pace at which research reports in behaviour change are published, researchers, consultants and policymakers would benefit from more automatic ways to process these reports. Automatic extraction of the rep…
How Far Are We from Genuinely Useful Deep Research Agents?
Deep Research Agents (DRAs) aim to automatically produce analyst-level reports through iterative information retrieval and synthesis. However, most existing DRAs were validated on question-answering benchmarks, while res…
Information RetrievalS1-DeepResearch: Beyond Search, Toward Real-World Long-Horizon Research Agents
Deep research agents aim to solve complex knowledge-intensive tasks through long-horizon planning, evidence gathering, reasoning, and report generation. While recent progress in search agents has demonstrated strong capa…
Instruction FollowingInformation RetrievalQuestion AnsweringTowards Trustworthy Report Generation: A Deep Research Agent with Progressive Confidence Estimation and Calibration
As agent-based systems continue to evolve, deep research agents are capable of automatically generating research-style reports across diverse domains. While these agents promise to streamline information synthesis and kn…
Mind2Report: Expert-Level Commercial Report Synthesis via Cognitive Deep Research Agent
Synthesizing informative commercial reports from massive and noisy web sources is critical for high-stakes business decisions. Although recent deep research agents (DRAs) achieve notable progress, their reports remain li…