Electimize: New Evolutionary Algorithm for Optimization with Application in Construction Engineering

    Authors

    A. Khalafallah;M. Abdel-Raheem

    Comments

    Authors: contact us about adding a copy of your work at STARS@ucf.edu

    Abbreviated Journal Title

    J. Comput. Civil. Eng.

    Keywords

    Optimization algorithms; Construction; Cash flow; GENETIC ALGORITHMS; PROJECTS; Computer Science, Interdisciplinary Applications; Engineering, Civil

    Abstract

    In nonlinear construction optimization problems, the capability of current optimization algorithms to find an optimal solution is usually limited by their inability to evaluate the effects of changing the value of each decision variable on reaching the optimal solution. This paper presents fundamental research aimed at developing a novel evolutionary optimization algorithm, named Electimize, that mimics the behavior of electrons flowing, through electric circuit branches with the least electric resistance. In the proposed algorithm, solutions are represented by electric wires and are evaluated on two levels: a global level, using the objective function, and a local level, evaluating the potential of each generated value for every decision variable. The paper presents (1) the research philosophy and scope, (2) the research methodology, and (3) the development of the algorithm. The proposed algorithm has been validated and applied successfully to an NP-hard cash flow optimization problem. The algorithm was able to find a better optimal solution and identified ten alternative optimal solutions for the same problem. This should prove useful in enhancing the optimization of complex large-scale problems. DOI: 10.1061/(ASCE)CP.19435487.0000080. (C) 2011 American Society of Civil Engineers.

    Journal Title

    Journal of Computing in Civil Engineering

    Volume

    25

    Issue/Number

    3

    Publication Date

    1-1-2011

    Document Type

    Article

    Language

    English

    First Page

    192

    Last Page

    201

    WOS Identifier

    WOS:000290451800003

    ISSN

    0887-3801

    Share

    COinS