Paper: arXiv 2311.16204
Authors: Tomás de la Rosa
Abstract
Once there is a decision of rebalancing or updating a portfolio of funds, the process of changing the current portfolio to the target one, involves a set of transactions that are susceptible of being optimized. This is particularly relevant when managers have to handle the implications of different types of instruments. In this work we present linear programming and heuristic search approaches that produce plans for executing the update. The evaluation of our proposals shows cost improvements over the compared based strategy. The models can be easily extended to other realistic scenarios in which a holistic portfolio management is required
Complexity vs Empirical Score
- Math Complexity: 6.5/10
- Empirical Rigor: 4.0/10
- Quadrant: Lab Rats — high math complexity, limited empirical testing
Why this score: The paper uses advanced math including linear programming, combinatorial optimization, and graph search algorithms, placing it in the high complexity category. While it presents evaluation of cost improvements, it lacks detailed backtesting metrics, statistical analysis, or real-world data implementation details, keeping the empirical rigor on the lower side.
Research Flowchart
flowchart TD A["Research Goal: Minimize transaction costs<br>when updating mutual fund portfolios"] --> B["Key Methodologies"] B --> B1["Linear Programming<br>Optimization"] B --> B2["Heuristic Search<br>Approach"] B1 --> C["Computational Process"] B2 --> C C["Apply models to real<br>portfolio rebalancing scenarios"] --> D["Key Findings"] D --> D1["Cost improvements over<br>standard rebalancing strategies"] D --> D2["Models easily extendable<br>to other portfolio management<br>scenarios"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes linear programming and heuristic search approaches to generate execution plans for updating mutual fund portfolios.
Source sentence
In this work we present linear programming and heuristic search approaches that produce plans for executing the update. The main contributions of this work are twofold: (1) formaliz-ing the portfolio update as a combinatorial problem, and (2) proposing the use of heuristic search to find the optimal up-date plan.
Assumes 2
- All portfolio holdings are mutual funds and all update operations are transfers from one fund to another (single transaction type). model
Source sentence
Mutual Fund Switches In the base case, we assume all portfolio holdings are mutual funds (i.e., they belong to the same type of instruments) and all update operations are transfers from one fund to another (i.e., single type of transaction).
- Additivity of linear programming programs allows pre‑selecting the cheapest option between direct switches and sell‑plus‑buy transactions. model
Source sentence
Under the additivity assumption of LP programs, this pro-posal is also equivalent (assuming non-equal costs) to pre-select the cheapest option between cij and c′ ij for each i and j and compute the solution in a single set of decision variables.
Data
- Random portfolio update problems synthetic unknown
Source sentence
To simulate diverse update tasks, we generated groups of random problems of incremental portfolio size, ranging from 4 up to 13 holdings. All portfolios were scaled to have 10k of money value.
Costs
Modeled
Variable fee of 1–10 basis points for all funds; ETFs have a fixed fee selected from 0.5, 1.0, or 2.5 (simulating exchange commissions). Mutual funds have no fixed fee.
Source sentence
Here we wanted the simulation to also include rebalance splits (e.g., 2% overweight in fund X is allocated 1% to fund Y and 1% to fund Z.) • ETFs have a fixed fee selected from the alternatives (0.5 1.0, 2.5) with the idea of simulating different exchange commissions. Mutual funds have no fixed fee for the sub-scription or redemption of shares. • All funds have a variable fee in basic points (bps) ranging from 1 to
Finds 5
- Both LP+ and A*‑fee achieved the optimal solution in all 180 random problems. performance
Source sentence
In 180 problems (20×9) both LP+ and Astar-fee achieved the optimal solution.
- The naive approach produced plans with higher cost across all portfolio sizes. performance
Source sentence
Naive approach is clearly pro-ducing plans of worse cost across the whole range of port-folios.
- DFBnB matched the optimal solution in 145 tasks, exploring the full search tree in only 74 of them. performance
Source sentence
The DFBnB best solution matched the optimal solution in 145 tasks, of which in only 74, the complete search tree was explored.
- LP+ matched the number of transactions of A*‑fee in only 148 plans, indicating limited awareness of plan length. performance
Source sentence
Regarding the plan length, LP+ only matched the num-ber of transactions with Astar-fee in 148 plans, confirming our initial conjecture that LP+ lacks awareness of the plan length.
- DFBnB produced three fee‑sub‑optimal plans that had fewer transactions than the optimal cost plans. performance
Source sentence
DFBnB produced 3 fee sub-optimal plans that have fewer number of transactions.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 20 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (22)
- Robust modeling of multi-stage portfolio problems (2000)
- Robust portfolio optimization: a categorized bibliographic review (2020)
- Online lazy updates for portfolio selection with transaction costs (2013)
- CVXPY : A P ython-embedded modeling language for convex optimization (2016)
- A rewriting system for convex optimization problems (2018)
- Artificial Intelligence: a modern approach (2009)
- Universal portfolios with and without transaction costs (1997)
- Hitchcock transportation problem (1956)
and 14 more
- Efficient algorithms for the Hitchcock transportation problem (1995)
- PDDL2. 1: An extension to PDDL for expressing temporal planning domains (2003)
- Temporal planning with preferences and time-dependent continuous costs (2012)
- A formal basis for the heuristic determination of minimum cost paths (1968)
- Branch and bound methods for the traveling salesman problem (1983)
- Online portfolio selection: A survey (2014)
- Optimal execution of portfolio transactions (2001)
- Robo-advisors and wealth management (2017)
- Robo-advisory: digitalization and automation of financial advisory (2018)
- Personalized Multiple Account Portfolio Optimization (2023)
- Best practices for portfolio rebalancing (2015)
- Portfolio optimization in practice (1992)
- Daily mutual fund flows and redemption policies (2007)
- Hedge fund diversification: How much is enough? (2002)
0 of 22 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.