We propose a computational framework to simulate different limbless locomotion strategies on frictional substrates. Organism body deformation is modelled with a multiplicative decomposition of the deformation gradient into active (growing) and passive (elastic) parts. Optimal locomotion is computed by solving a minimisation problem with constraints that correspond to a two-field formulation and FE discretised algebraic differential equations. Adjoint-based approach is employed to deduce the first-order necessary conditions. The resulting nonlinear programming problem is solved with the forward-backward sweep method (FBSM). Our numerical experiments indicate that in the absence of inertia and gravity forces, inching (caterpillar) gait is energetically most efficient compared to crawling (maggot) and undulatory (C elegans) strategies. Moreover, it has been observed that optimal inching gait approaches the crawling as the organism specific weight ($\gamma:=\rho g$, where $\rho$ is material density and $g=9.81$) exceeds a critical number ($\gamma > 0.05$).