Chromosome representation through adjacency matrix in evolutionary circuits synthesis

Evolutionary Optimization of Microwave Filters

“…In the representation presented in this chapter, the Positional Matrix is always associated with valid circuits, what does not occur in other methods, with always present the generation of anomalous circuits in each generation. In the representation proposed by (Mesquita et al, 2002) the percentual of anomalous circuits is about 5% , but the author says that this percentage can reach 80%, which is undesirable.…”

“…In (Mesquita et al, 2002) a matrix representation for circuit synthesis is proposed but the mapping bi to one-dimensional is performed before the crossover operator is applied. However, in (Im et al, 2003), it is shown that this procedure losses neighbor structures, which is fundamental for the good behavior of the evolution process.…”

Evolutionary synthesis of low-sensitivity digital filters using adjacency matrix

“…The adjacency matrix is a convenient representation of circuit topology in some network analysis problems [42] if self-loops, representing short-circuits in the analog circuits domain, are excluded from the definition. This is achieved by restating the Definition 1 as:…”

“…The efficiency of an evolutionary algorithm requires suitable chromosome coding schemes able to avoid or competently test the generation of anomalous individuals during the evolutionary process [42,43]. When dealing with digital filters, there are two possible anomalies: structurally invalid individuals and non-computable individuals containing delay-free loops [1].…”

“…In this work a novel chromosome coding scheme based on the graph adjacency matrix [42] to perform the evolutionary synthesis of low-sensitivity digital filters is proposed. It is shown that the proposed genetic operators (crossover and mutation) through adjacency matrix representation can efficiently move within the search space when compared with the GP-tree representation [22] used to the same problem.…”