For offline data-driven multiobjective optimization problems (MOPs), no new data is available during the optimization process. Approximation models (or surrogates) are first built using the provided offline data, and an optimizer, for example, a multiobjective evolutionary algorithm, can then be utilized to find Pareto optimal solutions to the problem with surrogates as objective functions. In contrast to online data-driven MOPs, these surrogates cannot be updated with new data and, hence, the approximation accuracy cannot be improved by considering new data during the optimization process. Gaussian process regression (GPR) models are widely used as surrogates because of their ability to provide uncertainty information. However, building GPRs becomes computationally expensive when the size of the dataset is large. Using sparse GPRs reduces the computational cost of building the surrogates. However, sparse GPRs are not tailored to solve offline data-driven MOPs, where good accuracy of the surrogates is needed near Pareto optimal solutions. Treed GPR (TGPR-MO) surrogates for offline data-driven MOPs with continuous decision variables are proposed in this paper. The proposed surrogates first split the decision space into subregions using regression trees and build GPRs sequentially in regions close to Pareto optimal solutions in the decision space to accurately approximate tradeoffs between the objective functions. TGPR-MO surrogates are computationally inexpensive because GPRs are built only in a smaller region of the decision space utilizing a subset of the data. The TGPR-MO surrogates were tested on distance-based visualizable problems with various data sizes, sampling strategies, numbers of objective functions, and decision variables. Experimental results showed that the TGPR-MO surrogates are computationally cheaper and can handle datasets of large size. Furthermore, TGPR-MO surrogates produced solutions closer to Pareto optimal solutions compared to full GPRs and sparse GPRs.

In the traffic light scheduling problem, the evaluation of candidate solutions requires the simulation of a process under various (traffic) scenarios. Thus, good solutions should not only achieve good objective function values, but they must be robust (low variance) across all different scenarios. Previous work has shown that combining IRACE with evolutionary operators is effective for this task due to the power of evolutionary operators in numerical optimization. In this article, we further explore the hybridization of evolutionary operators and the elitist iterated racing of IRACE for the simulation–optimization of traffic light programs. We review previous works from the literature to find the evolutionary operators performing the best when facing this problem to propose new hybrid algorithms. We evaluate our approach over a realistic case study derived from the traffic network of Málaga (Spain) with 275 traffic lights that should be scheduled optimally. The experimental analysis reveals that the hybrid algorithm comprising IRACE plus differential evolution offers statistically better results than the other algorithms when the budget of simulations is low. In contrast, IRACE performs better than the hybrids for a high simulations budget, although the optimization time is much longer.

A recent comparison of well-established multiobjective evolutionary algorithms (MOEAs) has helped better identify the current state-of-the-art by considering (i) parameter tuning through automatic configuration, (ii) a wide range of different setups, and (iii) various performance metrics. Here, we automatically devise MOEAs with verified state-of-the-art performance for multi- and many-objective continuous optimization. Our work is based on two main considerations. The first is that high-performing algorithms can be obtained from a configurable algorithmic framework in an automated way. The second is that multiple performance metrics may be required to guide this automatic design process. In the first part of this work, we extend our previously proposed algorithmic framework, increasing the number of MOEAs, underlying evolutionary algorithms, and search paradigms that it comprises. These components can be combined following a general MOEA template, and an automatic configuration method is used to instantiate high-performing MOEA designs that optimize a given performance metric and present state-of-the-art performance. In the second part, we propose a multiobjective formulation for the automatic MOEA design, which proves critical for the context of many-objective optimization due to the disagreement of established performance metrics. Our proposed formulation leads to an automatically designed MOEA that presents state-of-the-art performance according to a set of metrics, rather than a single one.

The configuration of algorithms is a laborious and difficult process. Thus, it is advisable to automate this task by using appropriate automatic configuration methods. The irace method is among the most widely used in the literature. By default, irace initializes its search process via uniform sampling of algorithm configurations. Although better initialization methods exist in the literature, the mixed-variable (numerical and categorical) nature of typical parameter spaces and the presence of conditional parameters make most of the methods not applicable in practice. Here, we present an improved initialization method that overcomes these limitations by employing concepts from the design and analysis of computer experiments with branching and nested factors. Our results show that this initialization method is not only better, in some scenarios, than the uniform sampling used by the current version of irace , but also better than other initialization methods present in other automatic configuration methods.

Research on multi-objective evolutionary algorithms (MOEAs) has produced over the past decades a large number of algorithms and a rich literature on performance assessment tools to evaluate and compare them. Yet, newly proposed MOEAs are typically compared against very few, often a decade older MOEAs. One reason for this apparent contradiction is the lack of a common baseline for comparison, with each subsequent study often devising its own experimental scenario, slightly different from other studies. As a result, the state of the art in MOEAs is a disputed topic. This article reports a systematic, comprehensive evaluation of a large number of MOEAs that covers a wide range of experimental scenarios. A novelty of this study is the separation between the higher-level algorithmic components related to multi-objective optimization (MO), which characterize each particular MOEA, and the underlying parameters—such as evolutionary operators, population size, etc.—whose configuration may be tuned for each scenario. Instead of relying on a common or “default” parameter configuration that may be low-performing for particular MOEAs or scenarios and unintentionally biased, we tune the parameters of each MOEA for each scenario using automatic algorithm configuration methods. Our results confirm some of the assumed knowledge in the field, while at the same time they provide new insights on the relative performance of MOEAs for many-objective problems. For example, under certain conditions, indicator-based MOEAs are more competitive for such problems than previously assumed. We also analyze problem-specific features affecting performance, the agreement between performance metrics, and the improvement of tuned configurations over the default configurations used in the literature. Finally, the data produced is made publicly available to motivate further analysis and a baseline for future comparisons.

Reducing the energy consumption of water distribution networks has never had more significance. The greatest energy savings can be obtained by carefully scheduling the operations of pumps. Schedules can be defined either implicitly, in terms of other elements of the network such as tank levels; or explicitly, by specifying the time during which each pump is on/off. The traditional representation of explicit schedules is a string of binary values with each bit representing pump on/off status during a particular time interval. In this paper, we formally define and analyze two new explicit representations based on time-controlled triggers, where the maximum number of pump switches is established beforehand and the schedule may contain fewer than the maximum number of switches. In these representations, a pump schedule is divided into a series of integers with each integer representing the number of hours for which a pump is active/inactive. This reduces the number of potential schedules compared to the binary representation, and allows the algorithm to operate on the feasible region of the search space. We propose evolutionary operators for these two new representations. The new representations and their corresponding operations are compared with the two most-used representations in pump scheduling, namely, binary representation and level-controlled triggers. A detailed statistical analysis of the results indicates which parameters have the greatest effect on the performance of evolutionary algorithms. The empirical results show that an evolutionary algorithm using the proposed representations is an improvement over the results obtained by a recent state of the art hybrid genetic algorithm for pump scheduling using level-controlled triggers.