Real-World Applications - Immunological Computation: Theory and Applications - page 192

A framework called multilevel immune learning algorithm (MILA) is proposed
by Dasgupta et al. (2003). h e novelty is twofold: (1) More mechanisms of natural
immune system are utilized in this algorithm, including T helper cells, T suppres-
sor, B cells, and antigen-presenting cells (APCs). h is makes it diff erent from other
models using only the concept of either B cells or T cells. (2) h is algorithm detects
in a multilevel, multiresolution fashion, making large space to explore e ciently for
anomaly-detection applications.

7.8 Solving Optimization Problems
Hajela and Lee (1996), Hajela et al. (1997), and Hajela and Yoo (1999) described
the implementation of a constrained genetic search to simulate the mechanics of an
immune-inspired algorithm to solve engineering-optimization problems.
Coello and Cortes (2002) used NS algorithm to handle the problem of infea-
sible solution. Accordingly, they assumed

= =

Feasible individuals Ab

Antigen-antibody interaction (clone mutate and increase a nity) used to
make infeasible individuals move to feasible solution space

Ag and infeasible individuals

Endoh et al. (1998) and Toma et al . (1999) proposed an adaptive optimization algo-
rithm for the traveling salesman problem. h is approach is based on the immune
network model and major histocompatibility complex (MHC) peptide presenta-
tion. Here, the immune network principles were used to simulate adaptive behavior
of agents, various concepts such as MHC to induce competitive behavior among
agents, T cells as control behavior, and B cells as produce behavior (Table 7.1 shows
the specifi c mappings and the schematics diagram [Figure 7.13] illustrates the cor-
responding components).

Table 7.1 Immune Cells and Molecules and Their Roles

in the n-TSP Problem Solving

Immune System Role in the n-TSP problem

Antigen Contains information about the cities

and salesmen

Macrophage Selects the city number that the

salesman agent must visit

T cells Help the activation of B cell

B cells Produce antibodies

Antibody Performs the behavior of an agent