Hybrid Optimization Model for Real-Time Pickup Delivery Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing route planning approaches for on-demand vehicle-based pickup and delivery services face challenges during peak demand times, leading to increased computational times and delays in generating digital pickup and delivery plans, and are inflexible in adapting to dynamic constraints such as time windows and vehicle capacity.
Innovation Solution
A dispatch system utilizing hybrid optimization processes, specifically combining constraint programming (CP) with mixed integer programming (MIP) and combinatorial relaxations, to efficiently generate high-quality pickup and delivery plans that enforce precedence relationships and adapt to dynamic constraints, reducing computational time to milliseconds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing route planning approaches are used during peak demand times, then service coverage is maintained, but computational time increases significantly causing delays in generating pickup and delivery plans
Solution Approach 1:
The planning horizon is divided into multiple time intervals, with different planning approaches applied to different segments. During peak demand intervals, a simplified model generates plans quickly, while during off-peak intervals, a more comprehensive model optimizes routes. This segmentation allows the system to maintain high productivity during peak times while still achieving optimization during less critical periods.
Solution Approach 2:
The system dynamically adjusts the planning approach based on real-time demand conditions. When demand exceeds a threshold indicating peak times, the system switches to a faster, simplified planning model. When demand is lower, it transitions to a more comprehensive optimization model. This dynamic adaptation resolves the contradiction by matching computational effort to actual system needs.
2Adaptability or versatility
If comprehensive route planning models are used to handle dynamic constraints, then adaptability to time windows and vehicle capacity improves, but computational complexity increases
Solution Approach 1:
The system introduces an intermediary layer that translates complex dynamic constraints into simplified parameters for the planning model. Time windows are converted into adjusted service times, and vehicle capacity constraints are transformed into modified routing rules. This intermediary representation maintains adaptability to dynamic constraints while reducing computational complexity.
Solution Approach 2:
The system changes key parameters of the planning model based on the type and severity of constraints. For example, when time windows are tight, the model adjusts service time parameters; when vehicle capacity is limited, it modifies load parameters. These parameter changes allow the system to adapt to different constraint scenarios without requiring a completely different computational model for each case.
Data Source
AI summary
Technologies for computing, within a real-time time interval, of a digital plan for a vehicle in response to receipt by an online system of a volume of short-route pickup-and-delivery requests that increases over an operational period, include: receiving, by a hybrid optimization model, a set of digital inputs; where the set of digital inputs includes (i) pairs of geo-location data associated with pickup-and-delivery requests and (ii) geo-location data associated with the vehicle and (iii) precedence data that indicate an order that geo-locations in a pair of geo-location data are to be visited; using the hybrid optimization model to generate, in an execution time that does not exceed the real-time time interval, a pickup and delivery plan that conforms to the precedence data as digital output in response to the set of digital inputs; where the hybrid optimization model includes a constraint programming-based solver that operates in coordination with at least one other solver; creating, digitally storing, transmitting, and causing display of the digital output on a display of a device that is associated with the vehicle.