Abstract With the rise of ready-to-assemble furniture, driven by international giants like IKEA, assembly services were increasingly offered by the same retailers. When planning orders with assembly services, the estimation of the service time leads to additional difficulties compared to standard delivery planning. Assembling large wardrobes or kitchens can take hours or even days while assembling a chair can be done in a few minutes. Combined with the usually vast amounts of offered products, a lot of knowledge is required to plan efficient and exact delivery routes. This paper shows how an artificial neural network (ANN) can be used to accurately predict the service time of a delivery based on factors such as the goods to be delivered or the personnel providing the service. The data used include not only deliveries with assembly of furniture, but also deliveries of goods without assembly and delivery of goods requiring electrical installation. The goal is to create a solution that can predict the time needed based on criteria such the type of furniture, the weight of the goods, and the experiences of the service technicians. The findings show that ANNs can be applied to this scenario and outperform more classical approaches, such as multiple linear regression or support vector machines. Still existing problems are largely due to the provided data, e.g., a large difference between the number of short and longer duration orders, which made it harder to accurately predict orders with longer duration.

Abstract Verifying the genuineness of official documents, such as bank checks, certificates, contract forms, bonds, etc., remains a challenging task when it comes to accuracy and robustness. Here, the genuineness is related to the degree of match of the signature contained in the documents relating to the original signatures of the authorized person. Signatures of authorized persons are considered known in advance. In this paper, a novel feature set is introduced based on quasi-straightness of boundary pixel runs for signature verification. We extract the quasi-straight line segments using elementary combinations of the directional codes from the signature boundary pixels and subsequently we obtain the feature set from various quasi-straight line classes. The quasi-straight line segments provide a blending of straightness and small curvatures resulting in a robust feature set for the verification of signatures. We have used Support Vector Machine (SVM) for classification and have shown results on standard signature datasets like CEDAR (Center of Excellence for Document Analysis and Recognition) and GPDS-100 (Grupo de Procesado Digital de la Senal). The results establish how the proposed method outperforms the existing state of the art.

This paper analyzes the multirobot sumo fight simulation. This simulation is based on a computational model of several sumo fighters, which physically interact while trying to move the opponent out of the arena (lost fight). The problem is optimized using a genetic algorithm (GA), where the capabilities of not only one particular robot but of all robots simultaneously are improved. In this particular problem setup, the problem definition changes depending on the optimization path, because all robots also get better, competing against each other. The influence of different operators of the GA is investigated and compared. This paper raises the questions, which genetically controlled capabilities (e.g. size, speed) are dominant over time and how they can be identified by a sensitivity analysis using a GA. The results shed light on which parameters are dominant. This experiment typically opens up interesting fields of further research, especially about how to address optimization problems, where the optimization process influences the search space and how to eliminate the factor of randomness.

The vast majority of user authentication in digital applications is based on alphanumeric passwords. Yet, due to severe problems that might arise with this approach, various efforts have been made in the last decade to replace this authentication paradigm. One candidate for the prospective paradigm shift might be found in the field of graphical passwords. The present paper introduces a novel framework for user authentication based on freehand sketches. The basic idea is that during the registration phase a user draws an arbitrary sketch in a specific drawing canvas (rather than typing a password). Registered users can then be authenticated whenever they are able to reproduce their personal sketch with sufficient precision. The major challenge of such a system is twofold. First, it has to provide a certain degree of error-tolerance such that the authentication of genuine users can be smoothly accomplished. Second, the system should detect even subtle forgeries and reject possible intruders. The main contributions of this paper are as follows. First, we formally represent the underlying sketches by means of strings and present a general authentication algorithm that is based on structural pattern recognition. Second, we present a novel cost model that is particularly useful in conjunction with string matching. Third, by means of an exhaustive empirical investigation using both random and skilled forgeries (stemming from several hundreds of users) we empirically confirm the feasibility of this particular authentication framework in a real-world scenario.

A previously proposed ACO game agent controller for steering the ghosts in the computer game Ms. Pac-Man has been improved by reengineering and extending the original controller with an adapted ACO algorithm. The number of free parameters is reduced by 30% and their values are optimized using a genetic algorithm. The proposed new ACO controller, named “fairGhosts”, is tested against other available game agent controllers within the existing Ms. Pac-Man simulator framework. This controller reaches a performance comparable to the upper third of all proposed controllers. In certain cases, it beats even the best controllers. Our “fairGhost” algorithm, its implementation and functionality are discussed.

In recent years, a trend towards alternative random or more precisely pseudorandom number generators could have been observed, viz. non-uniform generators that are based on chaotic maps (e.g. Ikeda Map) and uniform generators that are based on linear recurrence (e.g. Xorshift). Especially, chaotic maps have shown their superiority over canonical pseudorandom number generators in different heuristics solving both single- and multi-objective problems. However, the reasons for this superiority have not been completely unveiled and are probably based on the non-uniform distribution of the generated pseudorandom numbers. The aim of this research is therefore to investigate the influence of both uniform and non-uniform pseudorandom number generators on the convergence behaviour and computing time of a Genetic Algorithm (GA) applied to an order picking problem. A GA is a heuristics that imitates the natural selection process using stochastic methods for both the initial creation as well as the further evolution of the population. The influence of different non-uniform pseudorandom generators is therefore compared against uniform pseudorandom number generators by optimising the picking order of a discrete warehouse. As a result, Xorshift-based pseudorandom number generators are often better - though not significantly - with respect to the convergence for less complex test cases. On the other hand, the Ikeda Map is often significantly better for more complex test cases. However, uniform Xorshift-based pseudorandom number generators are often superior with respect to computation time.

Chaotic maps are an alternative for calculating pseudorandom numbers which have created an increased interest among researchers dealing with stochastic search and optimization algorithms in the recent past. This interest is based on promising results with respect to both the quality of the results as well as the running time of the optimization algorithms compared to the usually used standard pseudorandom number generators. In this paper we investigate the influence of nine different chaotic maps on the quality of the results obtained by a self-organizing map (SOM) which has been used to solve the Traveling Salesman Problem (TSP). The investigation is based on various sizes of both the problem instances as well as the number of iterations where all nine chaotic maps are compared against the pseudorandom number generation. As a result it is proven that chaotic maps are significantly better in several cases. Finally, possible reasons for both the superiority and inferiority of chaotic maps compared to pseudorandom number generation are analyzed and discussed.

In this paper we consider the problem of an optimal picking order sequence in a multi-aisle warehouse that is operated by a single automatic storage and retrieval system (AS/RS). The problem is solved by using a genetic algorithm (GA) similar to the one in the earlier research [3]. The problem and the solution approach are implemented in the OpenOpal software which provides a suitable test bed for simulation and optimization (see http://www.openopal.org/). As a result it becomes evident that the genetic algorithm can be improved by changing the selection method and introducing an elitism mechanism.

An optimization framework combines various methods, strategies, and programming interfaces on a robust software platform. Its development requires knowledge from application areas, and about optimization methods, as well as from software engineering. Different persons provide diverse know-how about modeling and simulating engineering and/or business problems, about search and optimization methods, and about new software trends to implement them into software. This paper describes the approach how an optimization framework based on evolutionary algorithms and other methods is developed in subsequent projects with application engineers and software developers cooperatively working together guaranteeing a sophisticated knowledge transfer. Therefore, particular knowledge management aspects are emphasized. As result, the optimization platform OpenOpal and the ideas behind its software architecture, supporting the know-how transfer, are presented. In order to continuously improve this optimization framework it is transferred into an open-source software initiative. The objective is to broaden the user group by increasing the number of knowledge contributors both from academia - integrating and testing newly developed optimization methods - and from various engineering areas - providing real-world problems to be solved.

The creation of staff rosters for emergency shifts for doctors is a complex task. To construct good rosters, many restrictions (e.g. holidays and workload) have to be taken into account. These restrictions have been mathematically specified for a concrete case in order to solve the problem afterwards with a straightforward genetic algorithm. Thereby the main focus lays on two different mutation methods and the combination of them. The results of this procedure will be discussed in this work.

In this paper we discuss a real-life problem in rail crew rostering. Specific emphasis is placed on the requirements of advanced approaches in rostering and the usage of optimization-based decision support. The modeling of the rostering problems is discussed including the treatment of constraints, the consideration of preferences, and the formulation of several objective functions. The specific solving method of the problem using an evolutionary algorithm and visualization and navigation tools for decision support are sketched briefly and some preliminary results are shown. Finally, some conclusions are presented.

We consider the problem of assigning train drivers to scheduled trains services, a combinatorial optimization problem which involves various hard and soft constraints. The problem is formulated as a single and a multiobjective optimization problem. A genetic algorithm is designed for solving it. As the problem is a real-life problem, various issues of application and utilization within a railway planning suite are discussed as well.

In this article, we present a novel evolutionary algorithm for approximating the efficient set of a multiobjective optimization problem (MOP) with continuous variables. The algorithm is based on populations of variable size and exploits new rules for selecting alternatives generated by mutation and recombination. A special feature of the algorithm is that it solves at the same time the original problem and a dual problem such that solutions converge towards the efficient border from two "sides", the feasible set and a subset of the infeasible set. Together with additional assumptions on the considered MOP and further specifications on the algorithm, theoretical results on the approximation quality and the convergence of both subpopulations, the feasible and the infeasible one, are derived.