[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-81579-en":3,"doc-seo-81579-105":30,"detail-sidebar-cat-0-en-105":84},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":11,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},81579,962075006959,"Anda","https://ap-avatar.wpscdn.com/avatar/e0002397efbe92a78e?_k=1776741047341049297",6,"Technology","Programming over Thinking: Efficient and Robust Multi-Constraint Planning","Multi-constraint planning requires selecting, refining, and validating candidate plans under multiple, possibly conflicting constraints, yet LLM-based approaches face inconsistency and escalating cost. Pure reasoning chains accumulate errors and become prohibitively expensive as constraints grow. Existing LLM+coding/solver methods sacrifice flexibility and reuse, often rebuilding code per query or relying on fixed solvers. SCOPE (Scalable COde Planning Engine) separates query-specific reasoning from generic execution, generating reusable deterministic solver functions, improving success rates while reducing inference latency and cost.","Programming over Thinking: Efficient and Robust Multi-Constraint Planning  \nDerrick Goh Xin Deik 1 , Quanyu Long 1 , Zhengyuan Liu2 , Nancy F. Chen2 , Wenya Wang 1 *  \n1Nanyang Technological University, Singapore  \n2Agency for Science, Technology and Research (A*STAR), Singapore  \n[gohx0043@e.ntu.edu.sg](gohx0043@e.ntu.edu.sg), [wangwy@ntu.edu.sg](wangwy@ntu.edu.sg)  \narXiv :2601 .09097v4 [ cs .AI] 10 Jul 2026  \nAbstract  \nMulti-constraint planning involves identifying, evaluating, and refining candidate plans while satisfying multiple, potentially conflicting constraints. Existing large language model (LLM) approaches face fundamental limitations in this domain. Pure reasoning paradigms, which rely on long natural language chains, are prone to inconsistency, error accumulation, and prohibitive cost as constraints compound. Conversely, LLMs combined with coding- or solver-based strategies lack flexibility: they often generate problem-specific code from scratch or depend on fixed solvers, failing to capture generalizable logic across diverse problems. To address these challenges, we introduce the Scalable COde Planning Engine (SCOPE), a framework that disentangles query-specific reasoning from generic code execution. By separating reasoning from execution, SCOPE produces solver functions that are consistent, deterministic, and reusable across queries while requiring only minimal changes to input parameters. SCOPE achieves state-of-the-art performance while lowering cost and latency. For example, with GPT- 4o, it reaches 93 . 1% success on TravelPlanner, a 61.6% gain over the best baseline (CoT) while cutting inference cost by 1.4x and time by 4.67x. Code is available at [https://github.com/DerrickGXD/SCOPE](https://github.com/DerrickGXD/SCOPE).  \n1 Introduction  \nPlanning is the process by which an agent organizes sequences of decisions or actions to achieve a goal. With the rapid advancement of LLMs, there has been growing interest in applying them to automate planning tasks (Wei et al., 2025) . Unlike traditional problem-solving tasks such as question answering or math reasoning, planning demands exploration of a rapidly expanding solution  \n* Corresponding author  \nspace, where small errors accumulate and lead to inconsistent or invalid outcomes. Recent development of LLMs in text-based reasoning (Wei et al., 2022 ; Yao et al., 2023a ; Yuan et al., 2025 ; Gui et al., 2025), powered by strong reasoning capabilities, has shown promise on simpler planning tasks. However, due to LLMs output being inherently probabilistic, these methods lack robustness, where small variations of reasoning paths can produce inconsistent or invalid solutions, and models often lose track of constraints or accumulate errors over long reasoning chains. In addition, they facescalability issues. Reasoning chains grow exponentially with task difficulty, require large numbers of tokens, and become increasingly costly to execute (refer to Figure 4) .  \nTo address these issues, prior work has explored integrating external solvers into the reasoning process (Chen et al., 2023 ; Jiang et al., 2024), leveraging their reliability and determinism to enforce constraints and verify solutions. Subsequent efforts have aimed to handle diverse constraints in planning (Hao et al., 2025b,a; Wang et al., 2025) . Existing methods either perform similar reasoning in natural language or generate similar solver code for each query. As a result, they do not support reusable execution abstractions and incur high computational cost. Nevertheless, existing code-based methods lack a systematic mechanism to capture the underlying logic of multi-constraint planning, resulting in inefficiency and increased error.  \nTo this end, we propose the Scalable COde Planning Engine (SCOPE), a multi-agent framework that employs a two-stage reasoning process to convert natural language queries into optimized structured representations, which are then processed by reusable solver functions. In t","cbCaie505kH0nZE6","https://ap.wps.com/l/cbCaie505kH0nZE6","pdf",5871250,1,53,"English","en",105,"# Introduction\n## Challenges in LLM-based planning\n## Integrating external solvers\n## Proposed approach: SCOPE","[{\"question\":\"How does SCOPE improve efficiency and reliability in planning?\",\"answer\":\"SCOPE disentangles query-specific reasoning from generic code execution: it converts queries into structured representations, then generates reusable deterministic solver functions. It also uses an automated parameter-free refinement step to enhance reliability without labor-intensive expert prompt design.\"}]","Programming over Thinking: Efficient and Robust Multi-Constraint Planning | PDF",1784174444,134,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":79,"head_meta":81,"extra_data":83,"updated_unix":28},"programming-over-thinking-efficient-and-robust-multi-constraint-planning","",{"@graph":36,"@context":78},[37,54,69],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/technology/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/programming-over-thinking-efficient-and-robust-multi-constraint-planning/81579/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-07-29","2026-07-16",true,{"@type":66,"interactionType":67,"userInteractionCount":11},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72],{"name":73,"@type":74,"acceptedAnswer":75},"How does SCOPE improve efficiency and reliability in planning?","Question",{"text":76,"@type":77},"SCOPE disentangles query-specific reasoning from generic code execution: it converts queries into structured representations, then generates reusable deterministic solver functions. It also uses an automated parameter-free refinement step to enhance reliability without labor-intensive expert prompt design.","Answer","https://schema.org",{"og:url":52,"og:type":80,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":82,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":85},[86,90,94,98,103,106,111,116,121,124,128],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":87,"show_sort_weight":88,"slug":89},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":91,"show_sort_weight":92,"slug":93},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Exam",70,"exam",{"id":99,"doc_module":4,"doc_module_name":46,"category_name":100,"show_sort_weight":101,"slug":102},5,"Comic",60,"comic",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":104,"slug":105},50,"technology",{"id":107,"doc_module":4,"doc_module_name":46,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":112,"doc_module":4,"doc_module_name":46,"category_name":113,"show_sort_weight":114,"slug":115},8,"Research & Report",30,"research-report",{"id":117,"doc_module":4,"doc_module_name":46,"category_name":118,"show_sort_weight":119,"slug":120},9,"Religion & Spirituality",20,"religion-spirituality",{"id":119,"doc_module":4,"doc_module_name":46,"category_name":122,"show_sort_weight":119,"slug":123},"World Cup","world-cup",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":126,"show_sort_weight":125,"slug":127},10,"Lifestyle","lifestyle",{"id":129,"doc_module":4,"doc_module_name":46,"category_name":130,"show_sort_weight":99,"slug":131},19,"General","general"]