Search results for " Computer Science"
showing 10 items of 3983 documents
Towards safe reinforcement-learning in industrial grid-warehousing
2020
Abstract Reinforcement learning has shown to be profoundly successful at learning optimal policies for simulated environments using distributed training with extensive compute capacity. Model-free reinforcement learning uses the notion of trial and error, where the error is a vital part of learning the agent to behave optimally. In mission-critical, real-world environments, there is little tolerance for failure and can cause damaging effects on humans and equipment. In these environments, current state-of-the-art reinforcement learning approaches are not sufficient to learn optimal control policies safely. On the other hand, model-based reinforcement learning tries to encode environment tra…
Evolution of OO Methods: the unified case
1997
This paper takes an evaluative look into OO methods and especially the evolution of the new snified method from its ancestors, OMT and OODA. The paper ries to classify the components of the earlier ethods and identify the parts that have been taken into the Unified ethod. The research applies the method metrics approach. For the sake of compactness we limit ourselves to the class diagram technique of all methods. We make observations about the number of concepts in each variation and show how the metrics can be used to analyse the changes in the techniques.
Supplier switching decisions
2007
Abstract In order to maximize profit, a buying firm should continuously search for and access sources which offer more favorable prices. While the literature is replete with works on the formation and development of buyer–supplier relationships, there is surprisingly only scarce research on the termination of such relationships and supplier switching. Using the concept of switching costs in a principal-agent framework, we at first analyze whether a firm switches the entire or a partial quantity to an alternative supplier when there is either symmetric or asymmetric information about the alternative supplier’s cost structure. Information asymmetry results in inert supplier switching decision…
A branch and bound algorithm for the matrix bandwidth minimization
2008
In this article, we first review previous exact approaches as well as theoretical contributions for the problem of reducing the bandwidth of a matrix. This problem consists of finding a permutation of the rows and columns of a given matrix which keeps the non-zero elements in a band that is as close as possible to the main diagonal. This NP-complete problem can also be formulated as a labeling of vertices on a graph, where edges are the non-zero elements of the corresponding symmetrical matrix. We propose a new branch and bound algorithm and new expressions for known lower bounds for this problem. Empirical results with a collection of previously reported instances indicate that the propose…
Asset and Liability Modelling for Participating Policies with Guarantee
2008
We study the problem of asset and liability management of participating insurance policies with guarantees. We develop a scenario optimization model for integrative asset and liability management, analyze the tradeoffs in structuring such policies, and study alternative choices in funding them. The nonlinearly constrained optimization model can be linearized through closed form solutions of the dynamic equations. Thus large-scale problems are solved with standard methods. We report on an empirical analysis of policies offered by Italian insurers. The optimized model results are in general agreement with current industry practices. However, some inefficiencies are identified and potential im…
Using PageRank for non-personalized default rankings in dynamic markets
2017
Abstract Default ranking algorithms are used to generate non-personalized product rankings for standard consumers, for example, on landing pages of online stores. Default rankings are created without any information about the consumers’ preferences. This paper proposes using the product centrality ranking algorithm (PCRA), which solves some problems of existing default ranking algorithms: Existing approaches either have low accuracy, because they rely on only one product attribute, or they are unable to estimate ranks for new or updated products, because they use past consumer behavior, such as previous sales or ratings. The PCRA uses the PageRank centrality of products in a product dominat…
Exact algorithms for the multi-compartment vehicle routing problem with flexible compartment sizes
2021
Abstract The multi-compartment vehicle routing problem with flexible compartment sizes is a variant of the classical vehicle routing problem in which customers demand different product types and the vehicle capacity can be separated into different compartments each dedicated to a specific product type. The size of each compartment is not fixed beforehand but the number of compartments is limited. We consider two variants for dividing the vehicle capacity: On the one hand the vehicle capacity can be discretely divided into compartments and on the other hand compartment sizes can be divided continuously. The objective is to minimize the total distance of all vehicle routes such that all custo…
Risky choices in strategic environments: An experimental investigation of a real options game
2019
Managers frequently make decisions under conditions of fundamental uncertainty due the stochastic nature of the outcomes and competitive rivalry. In this study, we experimentally test a theoretical model under fundamental uncertainty and competitive rivalry by designing a sequential interaction game between two players. The first mover can decide either to choose a sure outcome that assigns a risky outcome to the second mover or to pass the decision to the second mover. If the second player gets the chance to decide, she can choose between a sure outcome, conditioned by the assignment of a risky payoff to the first mover, or the sharing of the risky outcome with the first mover. We then int…
Efficiency in constrained continuous location
1998
Abstract We present a geometrical characterization of the efficient, weakly efficient and strictly efficient points for multi-objective location problems in presence of convex constraints and when distances are measured by an arbitrary norm. These results, established for a compact set of demand points, generalize similar characterizations previously obtained for uncontrained problems. They are used to show that, in planar problems, the set of constrained weakly efficient points always coincides with the closest projection of the set of unconstrained weakly efficient points onto the feasible set. This projection property which are known previously only for strictly convex norms, allows to e…
A GRASP heuristic for the mixed Chinese postman problem
2002
Abstract Arc routing problems (ARPs) consist of finding a traversal on a graph satisfying some conditions related to the links of the graph. In the Chinese postman problem (CPP) the aim is to find a minimum cost tour (closed walk) traversing all the links of the graph at least once. Both the Undirected CPP, where all the links are edges that can be traversed in both ways, and the Directed CPP, where all the links are arcs that must be traversed in a specified way, are known to be polynomially solvable. However, if we deal with a mixed graph (having edges and arcs), the problem turns out to be NP -hard. In this paper, we present a heuristic algorithm for this problem, the so-called Mixed CPP…