Abstract
The resource-constrained project scheduling problem (RCPSP) is a well-known scheduling problem that has attracted attention since several decades. Despite the rapid progress of exact and (meta-)heuristic procedures, the problem can still not be solved to optimality for many problem instances of relatively small size. Due to the known complexity, many researchers have proposed fast and efficient meta-heuristic solution procedures that can solve the problem to near optimality. Despite the excellent results obtained in the last decades, little is known why some heuristics perform better than others. However, if researchers better understood why some meta-heuristic procedures generate good solutions for some project instances while still falling short for others, this could lead to insights to improve these meta-heuristics, ultimately leading to stronger algorithms and better overall solution quality. In this study, a new hardness indicator is proposed to measure the difficulty of providing near-optimal solutions for meta-heuristic procedures. The new indicator is based on a new concept that uses the o-distance metric to describe the solution space of the problem instance, and relies on current knowledge for lower and upper bound calculations for problem instances from five known datasets in the literature. This new indicator, which will be called the o -D indicator, will be used not only to measure the hardness of existing project datasets, but also to generate a new benchmark dataset that can be used for future research purposes. The new dataset contains project instances with different values for the o -D indicator, and it will be shown that the value of the o-distance metric actually describes the difficulty of the project instances through two fast and efficient meta-heuristic procedures from the literature.

Abstract
The branch-and-bound (B&B) procedure is one of the most widely used techniques to get optimal so-lutions for the resource-constrained project scheduling problem (RCPSP). Recently, various components from the literature have been assembled by Coelho and Vanhoucke (2018) into a unified search algo-rithm using the best performing lower bounds, branching schemes, search strategies, and dominance rules. However, due to the high computational time, this procedure is only suitable to solve small to medium-sized problems. Moreover, despite its relatively good performance, not much is known about which components perform best, and how these components should be combined into a procedure to maximize chances to solve the problem. This paper introduces a structured prediction approach to rank various combinations of components (configurations) of the integrated B&B procedure. More specifically, two regression methods are used to map project indicators to a full ranking of configurations. The objec-tive is to provide preference information about the quality of different configurations to obtain the best possible solution. Using such models, the ranking of all configurations can be predicted, and these predic-tions are then used to get the best possible solution for a new project with known network and resource values. A computational experiment is conducted to verify the performance of this novel approach. Fur-thermore, the models are tested for 48 different configurations, and their robustness is investigated on datasets with different numbers of activities. The results show that the two models are very competitive, and both can generate significantly better results than any single-best configuration.

Abstract
In the past few years, the genetic programming approach (GP) has been successfully used by researchers to design priority rules for the resource-constrained project scheduling problem (RCPSP) thanks to its high generalization ability and superior performance. However, one of the main drawbacks of the GP is that the fitness evaluation in the training process often requires a very high computational effort. In order to reduce the runtime of the training process, this research proposed four different surrogate models for the RCPSP. The experiment results have verified the effectiveness and the performance of the proposed surrogate models. It is shown that they achieve similar performance as the original model with the same number of evaluations and better performance with the same runtime. We have also tested the performance of one of our surrogate models with seven different population sizes to show that the selected surrogate model achieves similar performance for each population size as the original model, even when the searching space is sufficiently explored. Furthermore, we have investigated the accuracy of our proposed surrogate models and the size of the rules they designed. The result reveals that all the proposed surrogate models have high accuracy, and sometimes the rules found by them have a smaller size compared with the original model.

Abstract
The resource-constrained project scheduling problem is a widely studied problem in the literature. The goal is to construct a schedule for a set of activities, such that precedence and resource constraints are respected and that an objective function is optimized. In project scheduling literature, summary measures are often used as a tool to evaluate the performance of algorithms and to analyze instances and datasets. They can be classified in two groups, network measures describe the precedence constraints of a project, while resource measures focus on the resource constraints of the instance. In this manuscript we make an exhaustive evaluation of the summary measures for project scheduling. We provide an overview of the most prevalent measures and also introduce some new ones. For our tests we combine different datasets from the literature and generate a new set with diverse characteristics. We evaluate the performance of the summary measures on three dimensions: consistency, instance complexity and algorithm selection. We conclude by providing an overview of which measures are best suited for each of the three investigated dimensions.

Abstract
In recent years, machine learning techniques, especially genetic programming (GP), have been a powerful approach for automated design of the priority rule-heuristics for the resource-constrained project scheduling problem (RCPSP). However, it requires intensive computing effort, carefully selected training data and appropriate assessment criteria. This research proposes a GP hyper-heuristic method with a duplicate removal technique to create new priority rules that outperform the traditional rules. The experiments have verified the efficiency of the proposed algorithm as compared to the standard GP approach. Furthermore, the impact of the training data selection and fitness evaluation have also been investigated. The results show that a compact training set can provide good output and existing evaluation methods are all usable for evolving efficient priority rules. The priority rules designed by the proposed approach are tested on extensive existing datasets and newly generated large projects with more than 1,000 activities. In order to achieve better performance on small-sized projects, we also develop a method to combine rules as efficient ensembles. Computational comparisons between GP-designed rules and traditional priority rules indicate the superiority and generalization capability of the proposed GP algorithm in solving the RCPSP.