{"paper_id":"2509.23557","title":"SIMPOL Model for Solving Continuous-Time Heterogeneous Agent Problems","authors":["Ricardo Alonzo Fernández Salguero"],"paper_date":"2025-09-28","paper_type":"theoretical","paper_url":"https://arxiv.org/pdf/2509.23557v1","page_url":"https://thequant.space/flowcharts/simpol-model-for-solving-continuous-time-heterogeneous-agent/","scores":{"empirical_rigor":3.5,"math_complexity":9.2,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes SIMPOL (Simplified Policy Iteration), a modular numerical framework that solves continuous-time heterogeneous agent models by iterating on consumption policies using Howard's policy iteration and an upwind finite difference scheme, with a novel post‑processing module to enforce re","quote":"This paper presents SIMPOL (Simplified Policy Iteration), a modular numerical framework for solving continuous-time heterogeneous agent models. A distinctive contribution is a novel consumption policy post-processing module that imposes regularity through smoothing and a projection onto an economically plausible slope band, improving convergence and model behavior.","summary":"This paper presents SIMPOL (Simplified Policy Iteration), a modular numerical framework for solving continuous-time heterogeneous agent models. The core economic problem, the optimization of consumption and savings under idiosyncratic uncertainty, is formulated as a coupled system of partial…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The upwind finite difference scheme is monotone, which is required for convergence to the correct viscosity solution.","type":"model","stated":"explicit","quote":"The *upwind* finite difference scheme is selected for its proven monotonicity property, which is crucial for convergence to the correct viscosity solution, as demonstrated by Barles and Souganidis (1991).","support":"full"},{"assumption":"Viscosity solutions exist and are unique for the HJB‑FPK system under very general conditions.","type":"model","stated":"explicit","quote":"The solution to these problems often requires the concept of viscosity solutions, a notion of a weak solution introduced by Crandall and Lions (1983) and extended to second-order equations by Lions (1983a,b), which guarantees existence and uniqueness under very general conditions.","support":"full"},{"assumption":"Consumption policies are initialized as a fraction of total income (c0(a)=k(ra+y)).","type":"execution","stated":"explicit","quote":"A common choice is a fraction of total income, c0(a) = k(ra+y). 2.","support":"full"},{"assumption":"The stationary wealth distribution satisfies zero‑flux (reflecting) boundary conditions at the domain edges.","type":"model","stated":"explicit","quote":"The condition that agents cannot leave the domain [0,Amax] imposes zero-flux (reflecting) boundary conditions: J(a) = µ∗(a)p(a)−1 2σ2p′(a) = 0 at a = 0 and a = Amax (22) Furthermore, the distribution must integrate to one: R Amax 0 p(a)da = 1.","support":"full"}]},"datasets":{"status":"none_found","items":[],"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":"SIMPOL’s diagnostics show robust convergence of the HJB solver, with residual and FOC error norms falling below tolerances after a finite number of iterations.","kind":"theoretical","quote":"able 3: Excerpt from the HJB Solver Convergence Trace Iteration HJB Residual (∞-norm) FOC Error (∞-norm) M-Matrix OK 1 2.729e+00 7.477e+00 Yes 50 8.694e-01 4.171e+00 Yes 100 1.411e+00 4.171e+00 Yes 200 8.694e-01 4.171e+00 Yes 325 1.431e+00 4.171e+00 Yes ... ... ... ... Final convergence is reached when both error norms fall below their respective tolerances.","support":"full"},{"finding":"The Wasserstein‑2 contraction diagnostic yields a median contraction ratio of 0.9888 (IQR 0.9884–0.9896), confirming the controlled diffusion process is a contraction and converges to a unique stationary distribution.","kind":"theoretical","quote":"Stability and Wasserstein-2 Contraction The W2 contraction diagnostic provides the strongest evidence of the model’s stability. The median ratio of the W2 distance after and before the evolution is 0.9888, with an interquartile range of (0.9884, 0.9896).","support":"full"},{"finding":"When σ=0, the SIMPOL solver reproduces the analytically known Merton optimal consumption policy with very low relative error, especially on fine grids.","kind":"theoretical","quote":"Validation with the Merton Model (σ = 0) The final test is the comparison with the Merton model without uncertainty. The SIMPOL solver, when run with σ = 0 and the option enforce mu nonneg sigma0=True, reproduces this policy with a very low relative error, especially on fine grids, confirming that the code correctly implements the economic logic.","support":"full"},{"finding":"The stationary distribution computed by SIMPOL conserves total mass (≈1) and yields zero net flux at the boundaries, matching Monte Carlo simulation results with low L1 error.","kind":"theoretical","quote":"The diagnostic results are excellent: the total mass of the distribution is practically 1, and the net fluxes at the domain boundaries are numerically zero. A comparison with a long-run Monte Carlo simulation (an alternative method for finding the stationary distribution) shows high similarity (low L1 error), validating the FPK solution.","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":17,"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":"2026-11-14T19:47:23Z"},"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/"}