Toward Optimal FDM Toolpath Planning with Monte Carlo Tree Search

Abstract

The most widely used methods for toolpath planning in fused deposition 3D printing slice the input model into successive 2D layers in order to construct the toolpath. Unfortunately slicing-based methods can incur a substantial amount of wasted motion (i.e., the extruder is moving while not printing), particularly when features of the model are spatially separated. In recent years we have introduced a new paradigm that characterizes the space of feasible toolpaths using a dependency graph on the input model, along with several algorithms to search this space for toolpaths that optimize objective functions such as wasted motion or print time. A natural question that arises is, under what circumstances can we efficiently compute an optimal toolpath? In this paper, we give an algorithm for computing fused deposition modeling (FDM) toolpaths that utilizes Monte Carlo Tree Search (MCTS), a powerful general-purpose method for navigating large search spaces that is guaranteed to converge to the optimal solution. Under reasonable assumptions on printer geometry that allow us to compress the dependency graph, our MCTS-based algorithm converges to find the optimal toolpath. We validate our algorithm on a dataset of 75 models and show it performs on par with our previous best local search-based algorithm in terms of toolpath quality. In prior work we speculated that the performance of local search was near optimal, and we examine in detail the properties of the models and MCTS executions that lead to better or worse results than local search.

Coordinated Toolpath Planning for Multi-Extruder Additive Manufacturing

“…Finally we are also exploring a more powerful algorithmic framework for multi-extruder toolpath planning. We recently applied a Monte Carlo tree search (MCTS) framework to toolpath planning [21] that makes use of the decentralized Monte Carlo Tree Search (Dec-MCTS) [27] algorithm to perform decentralized coordinated planning of each arm/extruder. Dec-MCTS is designed to find globally optimal solutions and has been shown to be far more computationally efficient than an equivalent centralized approach for coordinated planning problems.…”

“…In prior work [7,[19][20][21], we presented a sequence of algorithms that seek to minimize the wasted motion or fabrication time of the extrusion process for a given input model. Our key insight was to observe that fabrication in a layer-by-layer fashion induces motion that can be eliminated by fabricating multiple layers in different parts of the model asynchronously.…”

“…The coordinated toolpath planning problem is to find collective trajectory X such that the utilization of the extruders is maximized. This maximization objective is analogous to parallel scheduling and differs from our prior work in singleextruder toolpath planning [7,[19][20][21], in which we minimized wasted motion (i.e., extrusionless travel). The collective trajectory can be built incrementally as it is executed, as we do in the solution presented in this paper.…”

As-Continuous-As-Possible Extrusion Fabrication of Surface Models

“…Existing slicing algorithms divide the model as a stack of flat layers by geometric operations. [Lensgraf and Mettu 2016Yoo et al 2020] proposed a series of optimization algorithms to minimize the total extrusionless travel distance (wasted motion or print time) in the space of feasible toolpaths. Although this is only for flat layers, We share similar precedence constraints.…”

“…Restrict the printing to flat layers, this problem can also be expressed PC-TSP. Efforts in the literature have been made to approximated solutions on solving this problem to minimize the wasted motion and thus the overall fabrication time Mettu 2017, 2018;Yoo et al 2020].…”