{"paper_id":"2507.06345","title":"Reinforcement Learning for Trade Execution with Market Impact","authors":["Patrick Cheridito","Moritz Weiss"],"paper_date":"2025-07-08","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2507.06345v1","page_url":"https://thequant.space/flowcharts/reinforcement-learning-for-trade-execution-with-market-impac/","scores":{"empirical_rigor":6,"math_complexity":8.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a novel reinforcement learning framework for optimal trade execution using logistic‑normal distributions to model market and limit order allocations.","quote":"We introduce a new approach. Our main contributions are: • The development of a novel reinforcement learning algorithm that uses the logistic-normal distribution to model random allocations as distributions over the simplex.","summary":"In this paper, we introduce a novel reinforcement learning framework for optimal trade execution in a limit order book. We formulate the trade execution problem as a dynamic allocation task whose objective is the optimal placement of market and limit orders to maximize expected revenue. By…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The algorithm must sell a fixed integer number of lots M within a finite horizon T.","type":"execution","stated":"explicit","quote":"We consider a trade execution algorithm selling M ∈N lots"},{"assumption":"At each decision time the algorithm observes the full market state and takes an action based on the previous sub‑interval.","type":"model","stated":"explicit","quote":"We assume at each time tn, the algorithm takes into account what happened during the preceding subinterval (tn−1, tn], and takes an action an immediately after tn."},{"assumption":"The simulated market contains noise traders, tactical traders, and strategic traders that react to the algorithm’s orders.","type":"market","stated":"explicit","quote":"Numerical experiments show that the proposed method outperforms traditional benchmark strategies in simulated limit order book environments featuring noise traders submitting random orders, tactical traders responding to order book imbalances, and a strategic trader..."},{"assumption":"Direct and indirect market impact emerge naturally from interacting agents in the simulation.","type":"market","stated":"implied","quote":"simulations contain trading agents reacting to the algorithm’s order placement, causing direct and indirect market impact."}]},"datasets":{"status":"available","items":[{"name":"Simulated limit order book environments","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"Numerical experiments show that the proposed method outperforms traditional benchmark strategies in simulated limit order book environments featuring noise traders..."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"synthetic"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"Market impact is measured by the mid‑price drop after market or limit order placement of various sizes.","quote":"For market orders, we observe a sharp price drop immediately after the order is placed, followed by a slight recovery across all market environments.","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Market impact is measured by the mid‑price drop after market or limit order placement of various sizes.","card_quote":"For market orders, we observe a sharp price drop immediately after the order is placed, followed by a slight recovery across all market environments.","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"execution-cost","findings":[{"finding":"The logistic‑normal (LN) algorithm outperforms the Dirichlet (DR) algorithm and standard benchmarks across most simulated markets.","kind":"performance","quote":"We observe that across all three market simulations, the LN algorithm outperforms the DR algorithm, except in the market with noise traders and in the market with noise and tactical traders when 60 lots are traded..."},{"finding":"The LN algorithm converges to higher rewards than the DR algorithm, especially for 20‑lot cases and most 60‑lot cases.","kind":"performance","quote":"We observe that the rewards for the LN algorithm converge better than for the DR algorithm"},{"finding":"The algorithm outperforms competitive benchmarks (SL, TWAP) in realistic simulations that account for direct and indirect market impact.","kind":"performance","quote":"The algorithm outperforms competitive benchmarks in realistic simulations that account for direct and indirect market impact."}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":36,"citations":{"cited_by_in_corpus":4,"by_independent_authors":3,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":2,"unclassified":2},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":true,"code_url":"https://github.com/moritzweiss/rlte","note":null},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T16:37:23Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}