François Guertin

A Tabu Search Heuristic for the Vehicle Routing Problem with Soft Time Windows

This paper describes a tabu search heuristic for the vehicle routing problem with soft time windows. In this problem, lateness at customer locations is allowed although a penalty is incurred and added to the objective value. By adding large penalty values, the vehicle routing problem with hard time windows can be addressed as well. In the tabu search, a neighborhood of the current solution is created through an exchange procedure that swaps sequences of consecutive customers (or segments) between two routes. The tabu search also exploits an adaptive memory that contains the routes of the best previously visited solutions. New starting points for the tabu search are produced through a combination of routes taken from different solutions found in this memory. Many best-known solutions are reported on classical test problems.

Parallel Tabu Search for Real-Time Vehicle Routing and Dispatching

An abundant literature about vehicle routing and scheduling problems is available in the scientific community. However, a large fraction of this work deals with static problems where all data are known before the routes are constructed. Recent technological advances now create environments where decisions are taken quickly, using new or updated information about the current routing situation. This paper describes such a dynamic problem, motivated from courier service applications, where customer requests with soft time windows must be dispatched in real time to a fleet of vehicles in movement. A tabu search heuristic, initially designed for the static version of the problem, has been adapted to the dynamic case and implemented on a parallel platform to increase the computational effort. Numerical results are reported using different request arrival rates, and comparisons are established with other heuristic methods.

A PARALLEL TABU SEARCH HEURISTIC FOR THE VEHICLE ROUTING PROBLEM WITH TIME WINDOWS

The vehicle routing problem with time windows models many realistic applications in the context of distribution systems. In this paper, a parallel tabu search heuristic for solving this problem is developed and implemented on a network of workstations. Empirically, it is shown that parallelization of the original sequential algorithm does not reduce solution quality, for the same amount of computations, while providing substantial speed-ups in practice. Such speed-ups could be exploited to quickly produce high quality solutions when the time available for computing a solution is reduced, or to increase service quality by allowing the acceptance of new requests much later, as in transportation on demand systems.

A GENETIC ALGORITHM FOR VEHICLE ROUTING WITH BACKHAULING.

In this paper, a greedy route construction heuristic for a vehicle routing problem with backhauling is described. This heuristic inserts customers one by one into the routes using a fixed a priori ordering of customers. Then, a genetic algorithm is used to identify an ordering that produces good routes. Numerical comparisons are provided with an exact algorithm and with other heuristic approaches.

A Tabu Search Heuristic for Resource Management in Naval Warfare

Effective utilization of scarce resources, in particular weapon resources, is a prominent issue in naval anti-air warfare. In this paper, defence plans are constructed to guide the allocation and scheduling of different types of defence weapons against anti-ship missiles, subject to various physical and operational constraints. To reduce the frequency of replanning, decision trees are considered to explicitly account, in a probabilistic manner, for all possible outcomes of a particular action. A construction heuristic is first developed to generate an initial tree. A tabu search heuristic then improves this tree through the removal or addition of defence actions, followed by update operations aimed at maintaining the consistency. Numerical results obtained on scenarios with an increasing number of threats show that substantial improvements, in terms of survivability of the ship, can be obtained in reasonable computation times using tabu search.

Meta-Heuristics for a Class of Demand-Responsive Transit Systems

The demand-adaptive systems studied in this paper attempt to offer demand-responsive services within the framework of traditional scheduled bus transportation: Users call to request service between two given points and, in so doing, induce detours in the vehicle routes; at the same time, though, a given set of compulsory stops is always served according to a predefined schedule, regardless of the current set of active requests. The model developed to select requests and determine the routing of the vehicle yields a difficult formulation but with a special structure that may be used to develop efficient algorithms. In this paper, we develop, test, and compare several solution strategies for the single line-single vehicle problem that belong to two general meta-heuristic classes, memory-enhanced greedy randomized multistart constructive procedures, and tabu search methods. Hybrid meta-heuristics combining the two methods are also analyzed.

Adaptive Memory Programming for a Class of Demand Responsive Transit Systems

In this paper we discuss Demand Adaptive Systems (DAS) which are intended as a hybrid public transportation system that integrates traditional bus transportation and on demand service, DAS lines regularly serve a given set of compulsory stops according to a predefined schedule and regardless of current demand. Between a compulsory stop and the next, optional stops can be activated on demand. Vehicles have to be rerouted and scheduled in order to satisfy as many requests as possible, complying with passage-time constraints at compulsory stops. This paper provides a general description of DAS, and discusses potential applications and solution methods, emphasizing differences and analogies with classical Demand Responsive Systems. The particular mathematical structure of DAS requires innovative solution methods even when addressing its simplest version, the single vehicle, single line case. An efficient meta-heuristic algorithm based on adaptive memory ideas has been developed for this case. The method integrates sophisticated mathematical programming tools into a tabu search framework, taking advantage of the particular structure of the problem. The methodology is briefly discussed and experimental results are presented for the single line case. We show that the basic case can be efficiently solved, thus providing efficient algorithmic building blocks for more comprehensive approaches tackling the general case.

The Clustered Traveling Salesman Problem: A Genetic Approach

The Clustered Traveling Salesman Problem is an extension of the classical Traveling Salesman Problem, where the set vertices is partitioned into clusters. The goal is to find the shortest tour in such a way that all vertices of each cluster are visited contiguously. In this paper, a genetic algorithm is proposed to solve this problem. Computational results are reported on a set of Euclidean problems, and comparisons are provided with another recent heuristic.

Coupling a Greedy Route Construction Heuristic with a Genetic Algorithm for the Vehicle Routing Problem with Time Windows

Vehicle routing algorithms can be divided into three broad classes: route construction heuristics that “build” routes through the insertion of new customers, route improvement heuristics that modify the location of customers within the existing routes through exchange procedures, and composite heuristics that mix route construction and route improvement procedures. In this paper, a greedy route construction heuristic for the vehicle routing problem with time windows is described. This heuristic inserts customers one by one into the routes using a fixed a priori ordering of the customers. Then, a genetic algorithm is proposed to identify the ordering that produces the best routes.

A genetic algorithm for the clustered traveling salesman problem with a prespecified order on the clusters

The Clustered Traveling Salesman Problem is an extension of the classical Traveling Salesman Problem, where the set of vertices is partitioned into clusters. The goal is to find the shortest tour such that the clusters are visited in a prespecified order and all vertices within each cluster are visited contiguously. In this paper, a genetic algorithm is proposed to solve this problem. Computational results are reported on a set of Euclidean problems and a comparison is provided with a recent heuristic.