Core work packages – PigIT, University of Copenhagen

Core work packages

Work package B.1. Monitoring and forecasting of growth and feed consumption

WP leader: Nils Toft, UCPH LAS

Purpose: Growth and feed consumption in growing pigs are important markers for production efficiency and, furthermore, decreased growth rate and/or changed eating patterns are important indicators of health and welfare problems. The goal is to develop algorithms and models to estimate growth and welfare of piglets and criteria for early warning in case of problems.

Deliverables: (a) algorithms for online weight assessment of individual weaners by vision technology; (b) dynamic modeling of growth and feed consumption by state space models; (c) patterns reflecting poor health/welfare or low production efficiency are identified; (d) criteria for alarms are identified; (e) forecasting; (f) scientific publications.

Work packages B.2-B.4

These packages are similar in the sense that they address important welfare problems which also have a negative influence on productivity. The ambition of these work packages is to use on-line data to identify the problems and, particularly, behavioral patterns preceding the condition. Where relevant, the EVOP principle will be used to identify how production factors should be combined in order to minimize the risk of the condition.

Work package B.2. Detection of tail biting, intervention and decision about tail docking

WP leader: Lene Juul Pedersen, AU ANIS.

A special aspect of this work package is that tail biting interacts with the tactical decision of tail docking at batch level. It is, therefore, an example of an optimization problem involving decisions at different levels and with different time horizons.

Work package B.3. Detection of diarrhea and intervention

WP leader: Helle Halkjær Kristensen, UCPH LAS.

Work package B.4. Detection of undesired excretory behavior and intervention

WP leader: Heidi Andersen, AU ANIS.

Deliverables for each of the work packages B.2-B.4  

(a) algorithms for detection of indicators for coming outbreak of the problem; (b) quantification of welfare problems related to the condition; (c) intervention options; (d) algorithms for determination of optimal intervention strategy (e) scientific publications.

Work package B.5. Detection of the duration of different production phases in growing pigs

WP leader: Erik Jørgensen, AU ANIS

Purpose: When pigs are moved from one pen to the other during the production process, it usually takes some time before they are settled down. Typically, this phase will coincide with disease and behavioral disturbances as well as reduced productivity. Following this phase, a phase with the maximum productivity will start. However, at the end of the growing period, the resources will become scarce as the pigs grow and often again the productivity will decline. It is our ambition to develop methods for modeling the timing of events in the production cycle and detect the duration and effects of different production phases in order to provide an effective support system in relation to outbreak of behavioral and production problems.

Deliverables: (a) experience reports on the duration and effects of different productions phases; (b) scientific publications

Work package B.6. Optimal feeding and marketing policy in slaughter pigs

WP leader: Anders Ringgaard Kristensen, UCPH LAS

Purpose: The large pen systems used for the slaughter pigs in the project provide automatic on-line weight assessment from a mechanical system at individual pig level. The system includes an option for dynamic allocation of pigs to two or three different feeding areas with separate feed rations. Feed consumption will be known at group level. The intensive monitoring combined with the different feeding levels enable us to estimate a pen specific production function linking daily gain and feed intake. It is our goal to define an optimization model for the determination of optimal feeding and marketing of pigs.

Deliverables: (a) a generic modeling of the feeding-growth relation (production function); (b) automatic learning of production function parameters from data; (c) improved algorithms for optimal dynamic allocation of pigs to the different feeding areas (feeding rations), feeding level in groups (if possible) and marketing of individual pigs for slaughter; (d) scientific publications.