Table 1 While the space complexity of the generated DP schemes is always bounded by \(O(n^{tw})\) (Lemma 3), the run-time complexity of filling-up the DP tables \(C_{\boxslash }\) and \(C_{\boxtimes }\) depends on the choice of energy model. As for the table corresponding to a transitional bag X with indices I, the cost of filling it is \(O(n^{tw+1})\) irrespectively of the energy model
From: Automated design of dynamic programming schemes for RNA folding with pseudoknots
Energy model | Diagonal tables | Clique tables | Transitional tables |
|---|---|---|---|
\(C_{\boxslash }[i,j\vert S]\) | \(C_{\boxtimes }[i,i',j',j]\) | \(M_X[I_X]\) | |
BP-based model | \(O\left( n^{\vert S\vert +2}\right)\) | \(O\left( n^4\right)\) | \(O\left( n^{|I|}\right)\) |
BP+stacking | \(O\left( n^{|S|+2}\right)\) | \(O\left( n^4\right)\) | | |
Full Turner | \(O\left( n^{|S|+3}\right)\) | \(O\left( n^5\right)\) | | |