A Multi-Objective Optimization Cutting Method for Irregularly Shaped Fish Fillets

Irregularly shaped fish fillets make it difficult for conventional cutting methods to simultaneously control slice weight, maintain shape consistency, and reduce residual material. To address this problem, this study proposed a multi-objective optimization cutting method based on line-laser scanning and mathematical modeling. First, three-dimensional point-cloud data of fish fillets were acquired, and a weight–length distribution model was established to describe the mass distribution along the fillet length. Then, a constrained multi-objective cutting model was developed by taking slice weight as a constraint and minimizing both the variation in slice cross-sectional area and the residual material weight. An improved Marine Predators Algorithm combined with Deb’s feasibility rules was used to solve the cutting-position optimization problem. The results showed that the weight prediction accuracy of the developed distribution model reached 94.4–96.6%. Compared with the conventional sequential cutting method, the proposed method reduced the relative standard deviation of slice cross-sectional area to 8.900%, 11.112%, and 13.601% at target slice weights of 10 g, 15 g, and 20 g, respectively. Meanwhile, the average residual material weights were reduced to 0.581, 0.693, and 0.710 times the theoretical residual quantities, and the mean relative errors of slice weight were all below 8%. These findings indicate that the proposed method can improve slice shape consistency and raw material utilization while maintaining acceptable weight accuracy, providing a useful approach for intelligent and standardized fish fillet cutting.