Emanuele Manni

A Comparison of Anticipatory Algorithms for the Dynamic and Stochastic Traveling Salesman Problem

Advances in information technology and telecommunications, together with ever-growing amounts of data, offer opportunities for transportation companies to improve the quality of the service that they provide to their customers. This paper compares two methods motivated by the opportunity that the availability of data and technology gives to improve on current practice. In particular, the two solution approaches are explored in the context of a dynamic and stochastic routing problem in which a single, uncapacitated vehicle serves a set of known customers locations. One approach, sample-scenario planning, offers the potential for higher-quality solutions, but at the expense of greater computational effort. On the other hand, anticipatory insertion offers reduced computation and increased managerial ease, but with the potential for reduced solution quality due to restrictions on solution structure. Our results show that anticipatory insertion can often match the quality of sample-scenario planning, particularly when the degree of dynamism is low.

Real-Time Fleet Management At Ecourier Ltd

In this chapter we describe an innovative real-time fleet management system designed and implemented for eCourier Ltd (London, UK) for which patents are pending in the United States and elsewhere. This paper describes both the business challenges and benefits of the implementation of a real-time fleet management system (with reference to empirical metrics such as courier efficiency, service times, and financial data), as well as the theoretical and implementation challenges of constructing such a system. In short, the system dramatically reduces the requirements of human supervisors for fleet management, improves service and increases courier efficiency. We first illustrate the overall architecture, then depict the main algorithms, including the service territory zoning methodology, the travel time forecasting procedure and the job allocation heuristic

The impact of an efficient collection sites location on the zoning phase in municipal solid waste management.

In this paper, we study two decisional problems arising when planning the collection of solid waste, namely the location of collection sites (together with bin allocation) and the zoning of the service territory, and we assess the potential impact that an efficient location has on the subsequent zoning phase. We first propose both an exact and a heuristic approach to locate the unsorted waste collection bins in a residential town, and to decide the capacities and characteristics of the bins to be located at each collection site. A peculiar aspect we consider is that of taking into account the compatibility between the different types of bins when allocating them to collection areas. Moreover, we propose a fast and effective heuristic approach to identify homogeneous zones that can be served by a single collection vehicle. Computational results on data related to a real-life instance show that an efficient location is fundamental in achieving consistent monetary savings, as well as a reduced environmental impact. These reductions are the result of one vehicle less needed to perform the waste collection operations, and an overall traveled distance reduced by about 25% on the average.

Simultaneous personnel and vehicle shift scheduling in the waste management sector

Urban waste management is becoming an increasingly complex task, absorbing a huge amount of resources, and having a major environmental impact. The design of a waste management system consists in various activities, and one of these is related to the definition of shift schedules for both personnel and vehicles. This activity has a great incidence on the tactical and operational cost for companies. In this paper, we propose an integer programming model to find an optimal solution to the integrated problem. The aim is to determine optimal schedules at minimum cost. Moreover, we design a fast and effective heuristic to face large-size problems. Both approaches are tested on data from a real-world case in Southern Italy and compared to the current practice utilized by the company managing the service, showing that simultaneously solving these problems can lead to significant monetary savings.

MIP neighborhood synthesis through semantic feature extraction and automatic algorithm configuration

We propose an Automatic Neighborhood Design algorithm.The procedure relies on the extraction of semantic features from a MIP model.The algorithm is assessed on four well-known combinatorial optimization problems. The definition of a good neighborhood structure on the solution space is a key step when designing several types of heuristics for Mixed Integer Programming (MIP). Typically, in order to achieve efficiency in the search, the neighborhood structures need to be tailored not only to the specific problem but also to the peculiar distribution of the instances to be solved (reference instance population). Nowadays, this is done by human experts through a time-consuming process comprising: (a) problem analysis, (b) literature scouting and (c) experimentation. In this paper, we illustrate an Automatic Neighborhood Design algorithm that mimics steps (a) and (c). Firstly, the procedure extracts some semantic features from a MIP compact model. Secondly, these features are used to derive automatically some neighborhood design mechanisms. Finally, the proper mix of such mechanisms is sought through an automatic configuration phase performed on a training set representative of the reference instance population. When assessed on four well-known combinatorial optimization problems, our automatically-generated neighborhoods outperform state-of-the-art model-based neighborhoods with respect to both scalability and solution quality.

Topics in real-time fleet management

This is a summary of the author’s Ph.D. thesis, supervised by Gianpaolo Ghiani and Barrett W. Thomas and defended on 21 February 2008, at the Università degli Studi della Calabria. The thesis is written in English and is available from the author upon request. In this work, various tactical and operational issues concerning real-time fleet management are studied. First, we introduce the Dynamic and Stochastic Traveling Salesman Problem and propose an optimal policy through a Markov Decision Process as well as develop lower and upper bounds on the optimal policy cost. Then, we present several strategies for implementing a priori routes, and we identify situations in which the use of more involved a priori strategies can give some benefit. Next, we consider the Dynamic and Stochastic Vehicle Dispatching Problem with Pickups and Deliveries, for which we develop anticipatory algorithms that evaluate alternative solutions through a short-term demand sampling and a fully sequential procedure for indifference zone selection. Finally, we propose Approximated Neighborhood Evaluation procedures for the same-day Courier Shift Scheduling Problem, a tactical problem which amounts to minimizing the staffing cost subject to probabilistic service level requirements.

The lane covering problem with time windows

Abstract The Lane Covering Problem with Time Windows arises in the context of collaborative logistics. Given a set of lanes, it aims at finding a set of tours covering all lanes with the objective of minimizing the total travel cost. The purpose of this paper is to formulate a model for such a problem and to propose a heuristic approach based on Lagrangian relaxation for its solution. The behavior of this procedure is tested on a set of random instances.

Path and speed optimization for conflict-free pickup and delivery under time windows

This article introduces a variant of the conflict-free pickup and delivery problem with time windows in which speeds can be regulated. The problem arises in several areas of transportation and logistics including routing and scheduling of automated guided vehicles in port terminals and coordination of unmanned aerial vehicles in controlled airspace. A particular aspect of this problem is that at most one vehicle can traverse an arc of the transportation network at any time. The problem studied in this paper is to determine the vehicle paths and speeds on each arc of the path in such a way that no conflicts arise, the time windows are met, and the total energy consumption is minimized. A branch-and-bound algorithm is described in which a lower bound is obtained by solving a separable nonlinear problem in quadratic time. If the solution of the relaxation is not conflict free, a set of space-based and time-based branching constraints are generated to resolve the detected conflicts. Computational experiments ...

Shift Scheduling Problem in Same-Day Courier Industry

This paper deals with the problem of minimizing the staffing cost of a same-day courier company subject to service-level requirements. The problem has been modeled as an integer program with nonlinear probabilistic constraints. We have devised a heuristic procedure to explore the search space efficiently through an approximated neighborhood evaluation (ANE) model, relying on the estimation (via simulation) of a reduced number of parameters. Computational results show that, compared to traditional neighborhood search procedures, the ANE approach provides significant cost reductions in a typical same-day courier setting.

Neighborhood Synthesis from an Ensemble of MIP and CP Models

In this paper we describe a procedure that automatically synthesizes a neighborhood from an ensemble of Mixed Integer Programming (MIP) and/or Constraint Programming (CP) models. We move on from a recent paper by Adamo et al. (2015) in which a neighborhood structure is automatically designed from a (single) MIP model through a three-step approach: (1) a semantic feature extraction from the MIP model; (2) the derivation of neighborhood design mechanisms based on these features; (3) an automatic configuration phase to find the “proper mix” of such mechanisms taking into account the instance distribution. Here, we extend the previous work in order to generate a suitable neighborhood from an ensemble of MIP and/or CP models of a given combinatorial optimization problem. Computational results show relevant improvements over the approach considering a single model.