Catálogo de publicaciones - libros

Many real world projects can be represented by stochastic activity network (SAN) models, where the duration of some or all of the project activities are, at best, known in probabilistic sense. These models are known in the literature as PERT networks and they are analyzed using the PERT procedure. It has been known for many years that the Project Evaluation and Review Technique (PERT) provides inaccurate information about the project completion time. Quite often this inaccuracy is large enough to render such estimates as not helpful. As a result of this inaccuracy, many improvements since the introduction of PERT in 1959 have been developed. However, in spite of this inaccuracy and the many improvements, PERT procedure continues to be taught and presented in most text books on Project Management. This is due, perhaps, to its simplicity, and ease of its application. In this paper a new alternative is developed that addresses the issues of accuracy and practicality simultaneously in analyzing SAN models. The new procedure is based on a more accurate representation of the distribution function of the project completion time. We first show that the project completion time can in certain instances be accurately represented by a normal distribution, but in many other instances it can not. In these other instances we show that the project completion time can be more accurately represented by an extreme value distribution. Hence, SAN is first characterized as to when we can use the normal distribution and when we can use extreme value distribution. In the first case, PERT estimates are accurate and will be used; however, in the second case a new procedure is developed that is easy to use, and results in more accurate estimates of the project completion time and its statistics. In this paper examples are also provided that illustrate the above characterization of SANs and the accuracy and practicality of the new procedure.

The vast majority of the research efforts in project scheduling over the past several years have concentrated on the development of exact and heuristic procedures for the generation of a workable baseline schedule assuming complete information and a static and deterministic environment. During project execution, however, a project may be subject to considerable uncertainty. Proactive-reactive project scheduling deals with uncertainty by creating a baseline schedule that is as much as possible protected against disruptions and by deploying reactive scheduling procedures to revise or reoptimize this schedule when necessary. This chapter focuses on the main principles of proactive-reactive scheduling and dwells on schedule robustness and its measurements. A number of recently developed proactive and reactive scheduling heuristics are described and their working principles are illustrated on a problem example

In the chapter we consider two different scheduling models for stochastic network projects. Models of the first type consider several simultaneously realized stochastic network projects of PERT type. Resource scheduling models of the second type also cover PERT type projects, but with two different kinds of renewable resources:

In all types of models each projects’ activity utilizes several non-consumable related resources with fixed capacities, e.g. machines or manpower. For each operation, its duration is a random variable with given density function. The first problem centers on determining: in order to minimize the average total expenses of hiring and maintaining resources subject to the chance constraints.

For the second class of developed models the problem boils down: in order to minimize average total projects’ expenses subject to the chance constraint.

Problems of resource project scheduling are solved via simulation, in combination with a cyclic coordinate descent method and a knapsack resource reallocation model.

Due dates are an essential feature of real projects, but little effort has been made in studying the RCPSP with due dates in the activities. This paper tries to bridge this gap by studying two problems: the TardinessRCPSP, in which the objective is total tardiness minimization and the DeadlineRCPSP, in which the due dates are strict (deadlines) and the objective is makespan minimization. The first problem is NP-hard and the second is much harder, since finding a feasible solution is already NP-hard. This paper has three objectives: Firstly to compare the performance on both problems of well-known RCPSP heuristics - priority rules, sampling procedures and metaheuristics - with new versions we have developed that take due dates into consideration. Secondly, to present an instance generator that can generate instances with loose, medium, and tight due dates for computational study. And, finally, to adapt the technique of justification to deal with due dates and deadlines and to show its profitability.

This paper presents a branch-and-cut based exact solution algorithm for scheduling of projects with variable intensity activities connected by feeding precedence constraints the objective being to minimize the violation of resource constraints. Feeding precedence constraints allow some overlap in the execution of the connected activities and capture the flow of material or information between them. New polyhedral results are obtained and computational results are summarized.

In this chapter project scheduling problems with setup times are considered. Some practical applications justifying considering setups separately from activities are described. An extensive classification of setup times adapted from machine scheduling is proposed, including activity vs. class setup, separable vs. inseparable setup, as well as sequence-independent and sequence-dependent setup times. A new category of setup times - schedule-dependent ones - is discussed. The main part of the paper shows how to model setup times in the presence of particular project components, such as: precedence constraints, resource availability constraints, multiple resource units requests, multiple resources, etc. Some possible extensions of the presented models are given.

We review the most recent lower bounds for the makespan minimization variant of the Resource Constrained Project Scheduling Problem. Lower bounds are either based on straight relaxations of the problems (e.g., single machine, parallel machine relaxations) or on constraint programming and/or linear programming formulations of the problem.

The justification technique was introduced various decades ago for the resource-constrained project scheduling problem, although it has rarely been used with the problem. Justification is a simple and quick technique which when applied to schedules produces a new schedule that is, at most, as long as the original schedule — and often shorter. A recent article , showed that incorporating justification in heuristic algorithms can produce a substancial improvement in the results obtained. These results have motivated us to generalise this technique in order to study it in greater depth. This paper proposes distinct forms and generalisations for the justification technique and studies the relation existing among sets of obtainable schedules. The obtained results show that the proposed generalisations are worthwhile. Several computational tests have been performed to ascertain the impact of the generalisations on algorithmic efficiency.

We introduce a generalized model of the resource constrained project scheduling problem (RCPSP). It features that (i) the duration of an activity is not constant, but can vary in a specified range, and (ii) the objective is to minimize a convex function of time-lag costs, where a time-lag cost is charged according to the difference between the start/completion times of activities. These features achieve the flexibility of the model. It is known that, in the RCPSP, resource constraints can be replaced by some precedence constraints appropriately defined between the activities that require a common scarce resource. If we remove resource constraints by precedence constraints, our problem can be formulated as the dual problem of a minimum cost flow problem, and thus can be solved efficiently. Exploiting this property, we design a heuristic algorithm based on local search. We conducted computational experiments with benchmark instances to minimize the weighted earliness-tardiness costs, as well as instances in which activity-crashing or relaxation of temporal constraints are allowed. These results indicate the usefulness of our generalized RCPSP model and the proposed algorithm.

In the last few years several heuristic, metaheuristic and hybrid techniques have been developed to solve the Resource-Constrained Project Scheduling Problem (RCPSP). Most of them use the standard activity list representation, given that it seems to perform best in solving the RCPSP independently of the paradigm employed (genetic algorithms, tabu search, simulated annealing, ...). However, we have designed an innovative representation, one which has not been used before and which includes a lot of problem-specific knowledge. Based on that representation we have developed a new competitive and robust hybrid genetic algorithm, which uses genetic operators and an improvement mechanism specially designed to work on that representation and exploit, in a very efficient way, the information contained in it. We have compared this algorithm with the best algorithms published so far, using the standard benchmark of PSPLIB. The results show the excellent performance of our algorithm.