{"paper_id":"2411.05417","title":"Optimal reinsurance and investment via stochastic projected gradient method based on Malliavin calculus","authors":["Yuta Otsuki","Shotaro Yagishita"],"paper_date":"2024-11-08","paper_type":"methodological","paper_url":"https://arxiv.org/pdf/2411.05417v1","page_url":"https://thequant.space/flowcharts/optimal-reinsurance-and-investment-via-stochastic-projected/","scores":{"empirical_rigor":3,"math_complexity":9.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a stochastic projected gradient method combined with Malliavin calculus to obtain optimal reinsurance and investment strategies.","quote":"This paper proposes a new approach using the stochastic projected gradient method and Malliavin calculus for optimal reinsurance and investment strategies.","summary":"This paper proposes a new approach using the stochastic projected gradient method and Malliavin calculus for optimal reinsurance and investment strategies. Unlike traditional methodologies, we aim to optimize static investment and reinsurance strategies by directly minimizing the ruin probability.…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The gradient of the objective function is ν‑Hölder continuous on the feasible set.","type":"model","stated":"explicit","quote":"While Lipschitz continuity of the gradient is a standard assumption (e.g., [7, 8]), we relax this assumption to H¨older continuity. Assumption 4.1. Let ν ∈(0, 1]. ∇f is ν-H¨older continuous on X: there exists L > 0, ν ∈(0, 1] such that ∥∇f(y) −∇f(x)∥≤L∥y −x∥ν for any x, y ∈X. We consider the mini-batch stochastic projected gradient method (Algorithm 1) for the problem (19).","support":"full"},{"assumption":"Stochastic gradients are unbiased and have bounded second moment (E[G|F]=∇f, E[‖G‖²|F] ≤ σ²).","type":"statistical","stated":"explicit","quote":"The first assumption requires the stochastic gradients to be unbiased. For any non-negative integer k, i, we have (a) E[G(xk, ξk,i)|Fk] = ∇f(xk), (b) E[∥G(xk, ξk,i)∥2|Fk] ≤σ2, where σ > 0 and {Fk} is the natural filtration of {xk}.","support":"full"},{"assumption":"The optimization problem is constrained but the constraint set is simple, allowing an unbiased Malliavin‑calculus gradient estimator.","type":"model","stated":"explicit","quote":"Fortunately, the constraint is simple and using Malliavin calculus for the Poisson process [19] yields an unbiased estimator of the gradient of the objective function.","support":"full"}]},"datasets":{"status":"available","items":[{"name":"Monte Carlo simulation for ruin probability estimation","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"To estimate these ruin probabilities, we performed the Monte Carlo method with sample size 10, 000.","support":"full"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"synthetic"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"none","findings":[{"finding":"A convergence analysis is established for the mini‑batch stochastic projected gradient method under Hölder‑continuous gradients for general constrained nonconvex problems.","kind":"theoretical","quote":"6 Concluding Remarks This paper has proposed a stochastic projected gradient method combined with Malliavin calculus to find optimal investment and reinsurance strategies. Secondly, we have considered a general constrained nonconvex optimization problem under the H¨older condition and have established convergence results of the stochastic projected gradient method.","support":"full"},{"finding":"Numerical experiments demonstrate that the proposed method effectively reduces ruin probabilities, achieving medians comparable to or lower than those from the adjustment‑coefficient approach.","kind":"performance","quote":"In Figure 4, the medians of the ruin probabilities obtained using our proposed method and the ruin probabilities obtained by the adjustment coefficient approach are at the same level. On the other hand, as shown in Figures 5-6, we see that the median of the ruin probabilities obtained using our proposed method appeared to be lower than that of the adjustment coefficient approach.","support":"full"},{"finding":"The method yields static optimal strategies, which are more practical than dynamic HJB‑based strategies.","kind":"theoretical","quote":"The strategies obtained through the adjustment coefficient approach are static, where the decision variables are real numbers, as opposed to dynamic strategies, such as those in the HJB approach. Our approach is more practical than the HJB approach because the strategies in ours is static as well as that in the adjustment coefficient approach.","support":"full"}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":14,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"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":2,"extracted_at":"2026-11-14T19:54:06Z"},"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/"}