Consolidating the Crowd: How Optimization Redefines Baggage Delivery Logistics

TL;DR

Crowdsourcing has disrupted logistics, but point-to-point delivery remains inefficient. This research explores a hybrid model that introduces order consolidation into a crowdsourced baggage delivery network. By moving from a simple matching platform to an optimization-driven consolidator, the study demonstrates a potential 42% reduction in total travel time and a 50%+ reduction in vehicle requirements, creating a more sustainable and cost-competitive business model.

The "Matching" Trap: Why Current Crowdsourcing Struggles

Most crowdsourcing platforms act as passive intermediaries. A customer requests a delivery, and a driver accepts it. While this "on-demand" model is flexible, it creates a massive amount of "idle capacity." In the context of baggage delivery—where tourists often travel from the same hub (the airport) to similar clusters (popular hotels)—sending a single vehicle for every single bag is economically and environmentally wasteful.

The authors identify a critical gap: Centralized Decision-Making. Without a platform that consolidates orders before assigning them, the network remains a collection of inefficient, independent actors rather than a cohesive logistics system.

Methodology: From Surveys to Evolutionary Algorithms

The researchers utilized a three-pronged approach to validate their model:

1. Consumer Insights: A survey of 294 tourists at Hong Kong International Airport revealed that "Price" and "Value-Added Services" are the primary drivers for adoption.
2. Benchmarking: Comparison with traditional providers (like WFS) showed that crowdsourcing can offer flexibility, but only consolidation can drive prices down to the $50-$200 HKD range tourists expect.
3. Route Optimization: Using a "milk-run" approach—historically used for collecting milk from various farms—the authors modeled baggage delivery as a Vehicle Routing Problem (VRP).

The Optimization Architecture

To solve the complex problem of picking up bags from multiple hotels and delivering them to the airport (The "Many-to-One" problem), the team used an evolutionary solution search approach (Genetic Algorithm). They integrated real-world data via Google Maps APIs to ensure travel times were realistic.

Fig 1. Visualizing a consolidated route: One vehicle serves multiple hotel hubs, drastically reducing redundant trips to the airport.

Proof in the Numbers: Efficiency Gains

The results of the numerical experiments were striking. The study compared a standard "Point-to-Point" (P2P) model against their "Consolidation" model across different scenarios.

1. Airport to Hotel (Distribution)

In a typical 30-minute window with 14 requests:

• P2P Model: Required 14 vehicles.
• Consolidation Model: Required only 6 vehicles.
• Impact: Total travel time dropped from 631 minutes to 363 minutes.

2. Hotel to Airport (Collection)

This is traditionally harder to coordinate. Across 8 different demand scenarios, consolidation outperformed P2P every time.

Table 1. The Consolidation model consistently shaved off hundreds of minutes in travel time compared to point-to-point delivery.

Critical Insight: The "Win-Win" Constraint

The study highlights a vital "physical intuition": consolidation only works if the Maximum Travel Time is capped. If a route becomes too long (e.g., over 90 minutes), customer satisfaction drops. However, within that 90-minute window, the platform can:

• Lower prices for customers by sharing the "ride" across multiple orders.
• Increase revenue for drivers by allowing them to earn more per trip than they would on a single delivery.

Conclusion & Future Outlook

This paper serves as a blueprint for the next generation of crowdsourcing apps. The transition from Platform-as-a-Matchmaker to Platform-as-an-Optimizer is the key to surviving in high-density urban markets like Hong Kong.

Limitations: The study assumes the platform has "perfect information" (all orders known in advance). Future research should explore dynamic consolidation, where routes are updated in real-time as new "last-minute" baggage requests appear on the app.

The Bottom Line: Efficiency in crowdsourcing doesn't come from more drivers; it comes from smarter routes.