The price of commitment in online stochastic vehicle routing

Conference · Thu Jan 01 04:00:00 EST 2009

OSTI ID:956696

Bent, Russell W ^[1]; Van Hentenryck, Pascal ^[2]

---

1. Los Alamos National Laboratory
2. BROWN UNIV

This paper considers online stochastic multiple vehicle routing with time windows in which requests arrive dynamically and the goal is to maximize the number of serviced customers. Early work has focused on very flexible routing settings where the decision to assign a vehicle to a customer is delayed until a vehicle is actually deployed to the customer. Motivated by real applications that require stability in the decision making, this paper considers a setting where the decision to assign a customer request to a vehicle must be taken when that request is accepted. Experimental results suggest that this constraint severely degrades the performance of existing algorithms. However, the paper shows how the use of stochastic information for vehicle assignment and request acceptance improves decision quality considerably. Moreover, the use of resource augmentation quantifies precisely the cost of commitment in online vehicle routing.

• Research Organization: Los Alamos National Laboratory (LANL)
• Sponsoring Organization: DOE
• DOE Contract Number: AC52-06NA25396
• OSTI ID: 956696
• Report Number(s): LA-UR-09-00015; LA-UR-09-15
• Country of Publication: United States
• Language: English

Related Subjects

97 MATHEMATICS AND COMPUTING
ALGORITHMS
ALLOCATIONS
ARTIFICIAL INTELLIGENCE
DECISION MAKING
INFORMATION
MEETINGS
OPTIMIZATION
PERFORMANCE
RESOURCES
ROUTING
STABILITY
STOCHASTIC PROCESSES
VEHICLES