Online Proactive Multi-Task Assignment with Resource Availability Anticipation

Déborah Conforto Nedelmann,Jérôme Lacan,Caroline Chanel

DOI: https://doi.org/10.4204/EPTCS.391.3

2023-10-04

Abstract:With the emergence of services and online applications as taxi dispatching, crowdsourcing, package or food delivery, industrials and researchers are paying attention to the online multi-task assignment optimization field to quickly and efficiently met demands. In this context, this paper is interested in the multi-task assignment problem where multiple requests (e.g. tasks) arrive over time and must be dynamically matched to (mobile) agents. This optimization problem is known to be NP-hard. In order to treat this problem with a proactive mindset, we propose to use a receding-horizon approach to determine which resources (e.g. taxis, mobile agents, drones, robots) would be available within this (possibly dynamic) receding-horizon to meet the current set of requests (i.e. tasks) as good as possible. Contrarily to several works in this domain, we have chosen to make no assumption concerning future locations of requests. To achieve fast optimized online solutions in terms of costs and amount of allocated tasks, we have designed a genetic algorithm based on a fitness function integrating the traveled distance and the age of the requests. We compared our proactive multi-task assignment with resource availability anticipation approach with a classical reactive approach. The results obtained in two benchmark problems, one synthetic and another based on real data, show that our resource availability anticipation method can achieve better results in terms of costs (e.g. traveled distance) and amount of allocated tasks than reactive approaches while decreasing resources idle time.

Multiagent Systems,Performance

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to optimize task allocation by predicting resource availability in online multi - task allocation, so as to improve the efficiency of task allocation and reduce costs. Specifically, researchers are concerned with how to effectively match dynamic requests with mobile agents (such as taxis, drones, etc.) when they arrive. This problem is very challenging in practical applications because it is necessary to make rapid and efficient decisions without knowing future requests.

To deal with this NP - hard problem, the author proposes a method based on a rolling time window. This method can predict which resources will become available within a future time window and perform task allocation accordingly. Different from traditional reactive methods, this method can consider the availability of resources in advance without assuming the distribution of future requests, thus achieving better task allocation results. In addition, the author also designs a genetic algorithm to solve this optimization problem. This algorithm comprehensively considers factors such as travel distance and request age to achieve a balance between cost and the number of allocated tasks.

In general, the main contribution of this paper is to provide a new online multi - task allocation method. This method can not only reduce the idle time of resources but also maximize the number of allocated tasks while reducing the total travel distance. This provides important theoretical and technical support for fields such as taxi dispatching and package delivery.