Parallel Tabu search heuristics for the dynamic multi-vehicle dial-a-ride problem (original) (raw)

Improved Tabu Search Heuristics for Static Dial-A-Ride Problem: Faster and Better Convergence

ArXiv, 2018

Multi-vehicle routing has become increasingly important with the rapid development of autonomous vehicle technology. Dial-a-ride problem (DARP), a variant of vehicle routing problem (VRP), deals with the allocation of customer requests to vehicles, scheduling the pick-up and drop-off times and the sequence of serving those requests by ensuring high customer satisfaction with minimized travel cost. In this paper, we propose an improved tabu search (ITS) heuristic for static DARP with the objective of obtaining high-quality solutions in short time. Two new techniques, initialization heuristic, and time window adjustment are proposed to achieve faster convergence to the global optimum. Various numerical experiments are conducted for the proposed solution methodology using DARP test instances from the literature and the convergence speed up is validated.

An Improved Tabu Search Heuristic for Static Dial-A-Ride Problem

arXiv (Cornell University), 2018

Multi-vehicle routing has become increasingly important with the rapid development in autonomous vehicle technology. Dial-a-ride problem, a variant of vehicle routing problem (VRP), deals with the allocation of customer requests to vehicles, scheduling the pickup and drop-off times and the sequence of serving those requests by ensuring high customer satisfaction with minimized travel cost. In this paper, we propose an improved tabu search (ITS) heuristic for static dial-a-ride problem (DARP) with the objective of obtaining high quality solutions in short time. Two new techniques, construction heuristic and time window adjustment are proposed to achieve faster convergence to global optimum. Various numerical experiments are conducted for the proposed solution methodology using DARP test instances from the literature and the convergence speed up is validated.

Solving the on-line Dial-a-Ride Problem

The Dial-a-Ride is an emerging transport system, in which a fleet of vehicles, without fixed routes and schedules, carries people from the desired pickup point to the desired delivery point, during a pre-specified time interval. It can be modelled as an N P-hard routing and scheduling problem, with a suitable mixed integer programming formulation. The exact approaches to this problem are insufficient for real-life problems: time dependent network, requests received on-line, different objective functions. In this paper an algorithm to solve the on-line DaR problem is proposed. It inserts instantaneously a new request and then starts an off-line optimisation phase based on the solution of an assignment problem inserted in a granular tabu search method. The algorithm, tested on instances created ad hoc using the network of Milan, prove to be fast and effective. Moreover, the same algorithm proved to be effective in solving the off-line version too.

Heuristic Algorithms for the Single Vehicle Dial-A-Ride Problem

Journal of the Operations Research Society of Japan

A numbet of papets have been proposed for the approximate solution of the diaUa-ride routing problem, Tl]e object of this paper is to examine some of these heuristics, to introduce some new heuristics, and to compare tliese approximate algoritluns.

Dynamic programming based metaheuristics for the dial-a-ride problem

The organization of a specialized transportation system to perform transports for elderly and handicapped people is usually modeled as dial-aride problem. Users place transportation requests with specified pickup and delivery locations and times. The requests have to be completed under user inconvenience considerations by a specified fleet of vehicles. In the dial-a-ride problem, the aim is to minimize the total travel times respecting time windows, maximum user ride times, and vehicle restrictions. This paper presents an exact dynamic programming algorithm for the dial-a-ride problem and an heuristic subjected to the dynamic programming concept, which is able to cope with the curse of dimensionality by restricting the considered solution space. Then, a hybrid Large Neighborhood Search is proposed applying the dynamic programming based algorithm. The algorithms are tested on a given set of benchmark instances from the literature.

A HYBRID GREEDY RANDOMIZED ADAPTIVE SEARCH HEURISTIC TO SOLVE THE DIAL-A-RIDE PROBLEM

Asia-Pacific Journal of Operational Research, 2013

This paper presents a hybrid metaheuristic for solving the static dial-a-ride problem with heterogeneous vehicles and fixed costs. The hybridization combines a reactive greedy randomized adaptive search, used as outer scheme, with a tabu search heuristic in the local search phase. The algorithm is evaluated on well-known instances taken from the literature and on a set of randomly generated test problems, varying in the number of customers. Extensive computational results show the effectiveness of the hybrid approach in terms of trade-off between solution quality and computational time.

Local search heuristics for the probabilistic dial-a-ride problem

OR Spectrum, 2009

This paper describes an efficient neighborhood search procedure for the probabilistic dial-a-ride problem. This procedure requires O(n 4) computations compared to O(n 6) if computed from scratch. Two heuristics with this search procedure embedded are developed for solving the problem: a tabu search heuristic and a hybrid GRASP-tabu search heuristic. Computational results show that tabu search is superior over GRASP-tabu search on the probabilistic dial-a-ride problem, and by incorporating the search technique resulted in faster heuristics.

Variable neighborhood search for the dial-a-ride problem

Computers & Operations Research, 2010

In dial-a-ride problems passengers have to be transported between pre-specified pickup and delivery locations under user inconvenience considerations. The problem variant considered in this paper aims at minimizing total routing costs while respecting maximum route duration limits, time windows, and maximum user ride time limits. We propose a competitive variable neighborhood search-based heuristic, using three classes of neighborhoods. The first neighborhood class uses simple swap operations tailored to the dial-a-ride problem; the second neighborhood class is based on the ejection chain idea; and the third neighborhood class exploits the existence of arcs where the vehicle load is zero, giving rise to natural sequences of requests. We report new best results for 16 out of 20 benchmark instances. problem with last-in-first-out loading restrictions . To our best knowledge this is the first time VNS is applied to the static singleobjective DARP.

A rejected-reinsertion heuristic for the static Dial-A-Ride Problem

Transportation Research Part B: Methodological, 2007

We present here a new heuristic for the static multi-vehicle Dial-A-Ride Problem, which we call a rejected-reinsertion heuristic. Its main objective is to minimize the number of vehicles used to satisfy all the demand, subject to service quality constraints. Passenger deviation from desired time is minimized in the scheduling stage after the insertion position is determined. This method improves the basic parallel insertion heuristic in two aspects. First, a rejected-reinsertion operation is performed each time it is infeasible to insert a new request into the vehicle routes. Each assigned request close to the new request in time frame and geographic location is tentatively removed from its current vehicle and the new request is inserted into the best position in that vehicle route, followed by the reinsertion of the removed request elsewhere in the system. Of all available rejected-reinsertions, the least-cost one is then implemented. Second, an improvement procedure including trip reinsertion and trip exchange operations is implemented periodically. Two sets of problems are tested in a computational study. These show that the proposed heuristic achieves vehicle reductions of up to 17% over the parallel insertion heuristic and is very efficient computationally.