Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{209384,
        author = {vanshika shah and vedant joshi and vedansh arora and sparsh jain and soham ratnaparikhi and soumya gattani},
        title = {Policy-Constrained Portfolio Allocation Using Linear Programming: A Spreadsheet-Based Verification, Sensitivity Analysis, and Mean-Absolute-Deviation Extension},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {5},
        pages = {1609-1617},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=209384},
        abstract = {Portfolio selection becomes even more complicated when investment policy imposes some constraints on capital, risk, concentration and liquidity. LP model is useful in addressing those types of problems because the imposed restrictions may be stated as linear constraints. Another method for incorporating portfolio risk in the model without specifying the covariance matrix, but maintaining LP nature of the problem is the Konno-Yamazaki mean-absolute-deviation (MAD) method. In this paper, the spreadsheet implementation of the models is developed and feasibility check, duality and sensitivity analysis are used for explaining the optimal portfolio selection. The first model allocates $10 million portfolio between five assets under the budget, risk-score, equity concentration, liquidity and asset-specific constraints. The annual return for the obtained portfolio is $934,000, which corresponds to the portfolio yield of 9.34%, with $3.4 million allocated to growth equities, $3.1 million to corporate bonds and $3.5 million to government securities. Budget, risk score and maximum 35% weight in government securities are binding constraints, with shadow prices of $0.032 per additional dollar of capital and $0.012 per additional risk point correspondingly. Moreover, the infeasibility of an earlier proposed allocation of $970,000 that exceeds the risk ceiling by 4 million points is identified. An increase. Increasing the risk ceiling from 50 million to 55 million points increases the optimum to $985,000, although the original shadow price cannot be applied throughout that interval. A second exercise applies the linearised MAD formulation to twelve months of synthetic returns for four assets and compares its portfolios with Markowitz mean–variance solutions. At the same target return, the two approaches can produce materially different allocations, with a maximum weight-distance of 25 percentage points in the reported cases. Because the data are synthetic and the exercise is in-sample, the findings are methodological rather than evidence of investment performance. The workbook is intended as an auditable academic implementation rather than as investment advice.},
        keywords = {},
        month = {October},
        }