Resource allocation in a collaborative reforestation value chain: Optimisation with multi-objective models.

The reforestation value chain depends on the selection of qualified seeds supplied from various sources to ensure the successful growth, as each reforestation site has particular ecological parameters. The reforestation process usually involves many partners from different organisations, increasing the complexity of seed allocation. This research addresses seed allocation in a collaborative, make-to-order reforestation value chain. Using multi-objective optimisation models and considering different degrees of collaboration, it aims to find the most compatible seeds for each reforestation site so as to favour regeneration success. As a case study, the models are applied to the Quebec reforestation value chain which manages over 1450 seed lots and an annual production of 130 million seedlings. The process must consider two groups of partners: a seed center, and 18 nurseries. The lexicographic method is used to solve the models. Results show that an array of optimal solutions favouring reforestation success are possible by considering the main objective in each model. The second objective, integrating partners' objectives separately, modifies the initial solution significantly. Furthermore, when the objectives of both groups of partners are considered simultaneously, the proposed allocation differs depending on their priority, while the reforestation success objective does not deteriorate. The proposed set of models provide decision makers with a means to rapidly find a suitable seed allocation plan that favours reforestation success while considering partners satisfaction and existing bottlenecks in the value chain. This article contributes to the field by providing a sustainable seed allocation model favouring reforestation success covering the three pillars of sustainability.

Removing the thermal component from heart rate provides an accurate VO2 estimation in forest work.

Heart rate (HR) was monitored continuously in 41 forest workers performing brushcutting or tree planting work. 10-min seated rest periods were imposed during the workday to estimate the HR thermal component (ΔHRT) per Vogt et al. (1970, 1973). VO2 was measured using a portable gas analyzer during a morning submaximal step-test conducted at the work site, during a work bout over the course of the day (range: 9-74 min), and during an ensuing 10-min rest pause taken at the worksite. The VO2 estimated, from measured HR and from corrected HR (thermal component removed), were compared to VO2 measured during work and rest. Varied levels of HR thermal component (ΔHRTavg range: 0-38 bpm) originating from a wide range of ambient thermal conditions, thermal clothing insulation worn, and physical load exerted during work were observed. Using raw HR significantly overestimated measured work VO2 by 30% on average (range: 1%-64%). 74% of VO2 prediction error variance was explained by the HR thermal component. VO2 estimated from corrected HR, was not statistically different from measured VO2. Work VO2 can be estimated accurately in the presence of thermal stress using Vogt et al.'s method, which can be implemented easily by the practitioner with inexpensive instruments.