![]() | Only 14 pages are availabe for public view |
Abstract Bilevel linear programming is known to be strongly NP-hard (non-deterministic polynomial-time hard), and it has been proven that merely evaluating a solution for optimality is also NP-hard task. This gives us an idea about the kind of challenges offered by bilevel problems with complex (non-linear, non-convex, discontinuous etc.) objectives and constraint functions. An interest in bilevel programming has been driven by a number of new applications arising in different fields of optimization. Artificial Immune Systems (AIS) represent a field of biologically inspired computing that attempts to exploit theories, principles, and concepts of modern immunology to design immune system based applications to solve problems in science and engineering. In this thesis, a hybrid strategy that utilizes principles from classical optimization within an evolutionary algorithm to quickly approach a bilevel optimum. The proposed method is a bilevel evolutionary algorithm based on a modified Artificial Immune System (AIS). Modified AIS focus the search for affinity with high degree by developing new effective strategies for affinity measures. An adaptive mutation used for achieving the diversity of antibodies and avoiding premature convergence to escape from local optima. The proposed AIS algorithm provides an analytical solution to Stackelberg game problem. Finally, comparing the modified AIS with other algorithms to illustrate the efficiency and accuracy of the modified algorithm. |