{"id":"https://openalex.org/W7160630676","doi":"https://doi.org/10.48550/arxiv.2605.05701","title":"Inference-Time Budget Control for LLM Search Agents","display_name":"Inference-Time Budget Control for LLM Search Agents","publication_year":2026,"publication_date":"2026-05-07","ids":{"openalex":"https://openalex.org/W7160630676","doi":"https://doi.org/10.48550/arxiv.2605.05701"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.05701","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.05701","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.05701","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135720814","display_name":"Zhengru Fang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fang, Zhengru","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135702608","display_name":"Senkang Forest Hu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Senkang Forest","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135717290","display_name":"Zhonghao Chang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chang, Zhonghao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135661702","display_name":"Yu Guo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Yu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135662825","display_name":"Yihang Tao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tao, Yihang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135704522","display_name":"Hongyao Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Hongyao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102808478","display_name":"M. Ruan","orcid":"https://orcid.org/0009-0008-0458-5751"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ruan, Mengzhe","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135693217","display_name":"Jun Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Jun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5016290340","display_name":"Yuguang Fang","orcid":"https://orcid.org/0000-0002-1079-3871"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fang, Yuguang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":9,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10286","display_name":"Information Retrieval and Search Behavior","score":0.722599983215332,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10286","display_name":"Information Retrieval and Search Behavior","score":0.722599983215332,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T13274","display_name":"Expert finding and Q&A systems","score":0.13729999959468842,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.06960000097751617,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/commit","display_name":"Commit","score":0.77920001745224},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.5884000062942505},{"id":"https://openalex.org/keywords/control","display_name":"Control (management)","score":0.5523999929428101},{"id":"https://openalex.org/keywords/dual","display_name":"Dual (grammatical number)","score":0.5339000225067139},{"id":"https://openalex.org/keywords/budget-constraint","display_name":"Budget constraint","score":0.5178999900817871},{"id":"https://openalex.org/keywords/action","display_name":"Action (physics)","score":0.40400001406669617},{"id":"https://openalex.org/keywords/audit","display_name":"Audit","score":0.37709999084472656},{"id":"https://openalex.org/keywords/trajectory","display_name":"Trajectory","score":0.36500000953674316}],"concepts":[{"id":"https://openalex.org/C153180980","wikidata":"https://www.wikidata.org/wiki/Q19776675","display_name":"Commit","level":2,"score":0.77920001745224},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6694999933242798},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.5884000062942505},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.5523999929428101},{"id":"https://openalex.org/C2780980858","wikidata":"https://www.wikidata.org/wiki/Q110022","display_name":"Dual (grammatical number)","level":2,"score":0.5339000225067139},{"id":"https://openalex.org/C8505890","wikidata":"https://www.wikidata.org/wiki/Q605095","display_name":"Budget constraint","level":2,"score":0.5178999900817871},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.40400001406669617},{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.37709999084472656},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.36500000953674316},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.33730000257492065},{"id":"https://openalex.org/C4679612","wikidata":"https://www.wikidata.org/wiki/Q866298","display_name":"Aggregate (composite)","level":2,"score":0.3165999948978424},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.31220000982284546},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.3091000020503998},{"id":"https://openalex.org/C122637931","wikidata":"https://www.wikidata.org/wiki/Q118084","display_name":"Unit (ring theory)","level":2,"score":0.2939999997615814},{"id":"https://openalex.org/C42475967","wikidata":"https://www.wikidata.org/wiki/Q194292","display_name":"Operations research","level":1,"score":0.2879999876022339},{"id":"https://openalex.org/C97854310","wikidata":"https://www.wikidata.org/wiki/Q19541","display_name":"Search engine","level":2,"score":0.2831000089645386},{"id":"https://openalex.org/C203479927","wikidata":"https://www.wikidata.org/wiki/Q5165939","display_name":"Controller (irrigation)","level":2,"score":0.275299996137619},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.272599995136261}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.05701","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.05701","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.05701","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.05701","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"sustainable_development_goals":[{"score":0.4804515838623047,"id":"https://metadata.un.org/sdg/17","display_name":"Partnerships for the goals"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"LLM":[0,161,217],"search":[1,43,81,109,218,228],"agents":[2,219],"increasingly":[3],"rely":[4],"on":[5,19],"tools":[6],"at":[7],"inference":[8],"time,":[9,82],"but":[10,37],"their":[11],"trajectories":[12],"are":[13],"often":[14],"constrained":[15],"by":[16],"hard":[17,180],"limits":[18],"both":[20,222],"tool":[21],"calls":[22],"and":[23,51,72,111,115,124,141,163,229],"generated":[24],"tokens.":[25],"Under":[26],"such":[27],"dual":[28,113],"budgets,":[29],"better":[30],"answers":[31],"require":[32],"not":[33],"only":[34,143],"stronger":[35],"models,":[36],"also":[38],"explicit":[39],"control":[40,199,215],"over":[41,173],"which":[42],"action":[44,88],"should":[45,220],"receive":[46],"the":[47,53,107,134,145,167,178,193,203,231],"next":[48],"budget":[49,78,105,165,187,214,224],"unit":[50,104],"when":[52,144,202],"accumulated":[54],"evidence":[55],"is":[56,206,225,234],"sufficient":[57],"to":[58,119,149],"commit":[59],"a":[60,89,129,138,151],"final":[61,232],"answer.":[62],"We":[63],"study":[64],"this":[65,117],"problem":[66],"in":[67],"multi-hop":[68,157],"question":[69],"answering":[70],"(QA)":[71],"formulate":[73],"it":[74],"as":[75,95],"two-stage":[76],"inference-time":[77,213],"control.":[79],"At":[80],"our":[83],"controller":[84],"assigns":[85],"each":[86],"feasible":[87],"task-level":[90],"Value-of-Information":[91],"(VOI)":[92],"score,":[93],"defined":[94],"an":[96],"operational":[97],"estimate":[98],"of":[99],"marginal":[100],"task":[101],"value":[102],"per":[103],"under":[106,177],"current":[108],"state":[110],"remaining":[112],"budget,":[114],"uses":[116],"score":[118],"choose":[120],"among":[121],"retrieval,":[122],"decomposition,":[123],"answer":[125,136,233],"commitment.":[126],"After":[127],"search,":[128],"selective":[130],"evidence-grounded":[131],"finalizer":[132],"compares":[133],"trajectory":[135],"with":[137],"refined":[139],"candidate":[140],"rewrites":[142],"residual":[146],"error":[147],"appears":[148],"be":[150],"low-risk":[152],"answer-form":[153],"error.":[154],"Across":[155],"four":[156,164,174],"QA":[158],"benchmarks,":[159],"three":[160],"backbones,":[162],"levels,":[166],"method":[168],"yields":[169],"positive":[170],"aggregate":[171],"gains":[172],"audited":[175],"baselines":[176],"same":[179],"dual-budget":[181],"protocol.":[182],"Ablations":[183],"show":[184],"that":[185,212],"search-time":[186],"control,":[188],"especially":[189],"budget-dependent":[190],"penalty,":[191],"provides":[192],"main":[194],"performance":[195],"gain,":[196],"while":[197],"answer-time":[198],"helps":[200],"mainly":[201],"retrieval":[204],"path":[205],"already":[207],"adequate.":[208],"These":[209],"results":[210],"suggest":[211],"for":[216],"govern":[221],"how":[223,230],"spent":[226],"during":[227],"committed.":[235]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-05-09T00:00:00"}
