Digital Model-Driven Genetic Algorithm for Optimizing Layout and Task Allocation in Human-Robot Collaborative Assemblies

TL;DR

This work tackles pre-deployment optimization of human–robot collaborative work-cells by embedding LO, TA, TS, and MP considerations into a single digital-model framework. A bilevel genetic algorithm drives a leader that sets layout and task allocation and a follower that schedules tasks, with a digital simulator providing KPI-driven feedback $F$ to guide optimization. A case study with a UR5e cobot demonstrates cycle-time reductions and practical, safe layouts compared to a random baseline, illustrating the framework's potential for design-time decision-making. The approach offers a general, simulation-based pipeline that can be extended to multi-objective settings and dynamic planning for more robust pre-deployment evaluation.

Abstract

This paper addresses the optimization of human-robot collaborative work-cells before their physical deployment. Most of the times, such environments are designed based on the experience of the system integrators, often leading to sub-optimal solutions. Accurate simulators of the robotic cell, accounting for the presence of the human as well, are available today and can be used in the pre-deployment. We propose an iterative optimization scheme where a digital model of the work-cell is updated based on a genetic algorithm. The methodology focuses on the layout optimization and task allocation, encoding both the problems simultaneously in the design variables handled by the genetic algorithm, while the task scheduling problem depends on the result of the upper-level one. The final solution balances conflicting objectives in the fitness function and is validated to show the impact of the objectives with respect to a baseline, which represents possible initial choices selected based on the human judgment.

Figures (5)

• Figure 1: (a) Final layout: $r_1$, $r_9$ and $r_{10}$ are fixed in the use case, while the other resources can be moved on the table. (b) Non-overlapping condition for the resources: in this case, the set $\bar{\mathcal{N}}=\{\hat{\textbf{n}}_q^1, \hat{\textbf{n}}_q^3\}$, therefore the constraint (\ref{['equation:upper_level_con6']}) is verified.
• Figure 2: Optimization vector for the leader problem.
• Figure 3: Block diagram of the optimization framework: every block contains the function that is called in the recursive process, while the continuous arrows represent the data flow.
• Figure 4: (a) Optimal scheduling corresponding to the best solution in Figure \ref{['fig:optimal_layout']}a. (b) Precedence graph of the emergency stop button: we assume that $o_{9-12}$ can only be executed by the human, while $o_{13}$ is collaborative.
• Figure 5: Data distributions for the metrics considered in the test, compared to the baseline. The labels represent the minimum, maximum and median values.