Box Allocation Optimization in Meal Kit Delivery
Box Allocation Optimization in Meal Kit Delivery
This study introduces the Box Allocation Problem (BAP), a novel optimization challenge in the $1.4 billion UK meal kit delivery market. BAP involves assigning orders across multiple production facilities to minimize daily recipe variations while adhering to capacity and eligibility constraints over a 15-day planning horizon. We formulate BAP as a mixed-integer linear programming (MILP) problem and systematically compare the performance of the COIN-OR Branch and Cut (CBC) solver with heuristic methods, including Tabu Search and Iterative Targeted Pairwise Swap. Scalability experiment on instances with up to 100,000 orders show that CBC consistently achieves optimal solutions in under two minutes, maintaining optimality even under dynamic conditions with fluctuating factory capacities and changing customer orders. By reducing day-to-day recipe discrepancies, this approach supports more accurate ingredient forecasting, decreases food waste, and improves operational efficiency across multi-factory network. These results provide the first comprehensive solution framework for temporal allocation problems in meal kit delivery operations.
math.OC
Nguyen, Thi Minh Thu
50464540-2387-4da8-852f-f0137356ba69
Genest, Loic
37d8ffe3-9b35-467c-a8ff-730f0e0da362
Zemkoho, Alain
30c79e30-9879-48bd-8d0b-e2fbbc01269e
Nguyen, Thi Minh Thu
50464540-2387-4da8-852f-f0137356ba69
Genest, Loic
37d8ffe3-9b35-467c-a8ff-730f0e0da362
Zemkoho, Alain
30c79e30-9879-48bd-8d0b-e2fbbc01269e
[Unknown type: UNSPECIFIED]
Abstract
This study introduces the Box Allocation Problem (BAP), a novel optimization challenge in the $1.4 billion UK meal kit delivery market. BAP involves assigning orders across multiple production facilities to minimize daily recipe variations while adhering to capacity and eligibility constraints over a 15-day planning horizon. We formulate BAP as a mixed-integer linear programming (MILP) problem and systematically compare the performance of the COIN-OR Branch and Cut (CBC) solver with heuristic methods, including Tabu Search and Iterative Targeted Pairwise Swap. Scalability experiment on instances with up to 100,000 orders show that CBC consistently achieves optimal solutions in under two minutes, maintaining optimality even under dynamic conditions with fluctuating factory capacities and changing customer orders. By reducing day-to-day recipe discrepancies, this approach supports more accurate ingredient forecasting, decreases food waste, and improves operational efficiency across multi-factory network. These results provide the first comprehensive solution framework for temporal allocation problems in meal kit delivery operations.
Text
2509.06157v1
- Accepted Manuscript
More information
Accepted/In Press date: 7 September 2025
Additional Information:
15 pages, 21 figures, 6 tables, submitted as a research article; September 2025 version
Keywords:
math.OC
Identifiers
Local EPrints ID: 508594
URI: http://eprints.soton.ac.uk/id/eprint/508594
ISSN: 2331-8422
PURE UUID: f8c0aa23-00e0-4024-b111-5b4562184ebd
Catalogue record
Date deposited: 27 Jan 2026 18:08
Last modified: 28 Jan 2026 03:37
Export record
Altmetrics
Contributors
Author:
Thi Minh Thu Nguyen
Author:
Loic Genest
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
