Programming Problem Solving Techniques

Programming Problem Solving Techniques-88
For example, the Traveling Salesman Problem is an example of a combinatorial optimization problem.

For example, the Traveling Salesman Problem is an example of a combinatorial optimization problem.A naive (such a Random Search) black box method may simply explore permutations of the cities.

Tags: Dissertation Health EconomicsEssay On Book A Faithful FriendJacksonian Democracy Thesis StatementThesis Construction SafetyLife Changing Decision EssaySolve My Pre Calculus Problem

From a Computational Intelligence perspective, one may consider the architecture, processes, and constraints of a given strategy as the features of an approach.

The suitability of the application of a particular approach to a problem takes into considerations concerns such as the (ability to address unexpected or unintended effects).

A of a system (tool, strategy, model) or a problem is a distinctive element or property that may be used to differentiate it from similar and/or related cases.

Examples may include functional concerns such as: processes, data structures, architectures, and constraints, as well as emergent concerns that may have a more subjective quality such as general behaviors, organizations, and higher-order structures.

A global search strategy provides the benefit of making few if any assumptions about where promising areas of the search space may be, potentially highlighting unintuitive combinations of parameters.

A local search strategy provides the benefit of focus and refinement of an existing candidate solution.The field of Data Mining has clear methodologies that guide a practitioner to solve problems, such as Knowledge Discovery in Databases (KDD) [Fayyad1996].Metaheuristics and Computational Intelligence algorithms have no such methodology.This section summarizes a general methodology toward addressing the problem of suitability in the context of Computational Intelligence tools.This methodology involves 1) the systematic elicitation of system and problem features, and 2) the consideration of the overlap of problem-problem, algorithm-algorithm, and problem-algorithm overlap of feature sets.Both of the above strategies suggest an iterative methodology, where the product or knowledge gained from one technique may be used to prime a subsequent stronger or weaker technique.An algorithm may be considered a strategy for problem solving.The stronger the method, the more that must be known about the problem domain.Rather than discriminating techniques into weak and strong it is more useful to consider a continuum of methods from pure block box techniques that have few assumptions about the problem domain, to strong methods that exploit most or all of the problem specific information available.Global is differentiated from Local Optimization in that the latter focuses on locating an optimal structure within a constrained region of the decision variable search space, such as a single peak or valley (basin of attraction).In the literature, global optimization problems refers to the class of optimization problems that generally cannot be addressed through more conventional approaches such as gradient descent methods (that require mathematical derivatives) and pattern search (that can get 'stuck' in local optima and never converge) [Price1977] [Toern1999].


Comments Programming Problem Solving Techniques

The Latest from ©