Self-organising virtual models to build cold-chain solutions

Using self-organising virtual models to plan, prioritise and scenario-test cold chain network deployment in the Global South

9th Oct 2024

By Dr Adam Gripton

As climate change introduces unprecedented uncertainty and ever-greater threats to the resilience of the global food supply chain, it has become more important than ever to consider ways in which Global South economies responsible for growing and harvesting primary crops can leverage the benefits of access to a cold-chain to reduce food waste and improve the security and reliability of harvested crop yields. This is particularly pertinent in countries where, historically, cold-chain facilities are seldom used by farmers to preserve the shelf-life of their produce. This frequently also exacerbates poverty levels, since smallholder farmers without access to reliable cooling facilities often have few options with each season’s harvest other than to sell their produce immediately locally or to an opportunistic trader at a low price-point, with the alternative being the spoilage of the harvest. Scaled up, this can hold such a country’s whole rural economy back, as access to lucrative export and higher value markets is contingent on being able to deliver and prove the cold-chain provenance of a product’s journey, limiting the ability of the typical smallholder farmer to build better outcomes for themselves through planning and product choice.

Building a new national network of cold-chain facilities can be attractive for government policymakers in the Global South as it can have impacts in several key areas: first, such facilities can provide wide-ranging improvements in profitability for agricultural producers in-country; second, the reduction in food loss that cold-chain facilities can provide can help achieve population nutrition targets; third, such cooling facilities may have the potential to house refrigerated medicine stock for national vaccination initiatives. Given the uncertain global market environment, it is important to be able to plan such a network deployment as a matter of critical national infrastructure, in much the same ways as a new road, rail or energy network might be planned. it is also important to be able to prioritise such a roll-out so that the initial deployments can demonstrate the most added value, ideally precipitating a domino effect when such value has been demonstrated to others within the country.

In order to assist in policymaking decisions, Clean Cooling Network (CCN) has developed a unique virtual simulation model of cold-chain operations that is linked to seasonal harvest patterns, economic and environmental conditions, transport network definition, intra-national demands and import and export points, and is capable of simulating the effects of deployment of facilities on the choices made by farmers: both in terms of the storage and customer for their harvested produce, but also in terms of the choices of product to cultivate on their land. This helps de-risk investment by providing a modelling environment where thousands of potential scenarios could be tested before anything is built, guiding policymakers towards the most important locations to prioritise for cold-chain facility deployment while also raising awareness of how such deployments are likely to change local economic patterns, enabling such effects to be planned for at an early stage.

Agent-based modelling for cold-chain operations

CCN’s virtual modelling environment, developed in-house using the MILES platform at the Centre for Sustainable Road Freight (SRF) at Heriot-Watt University , seeks to provide simulation-based insights into the complex, multi-layered behaviours inherent within food supply chain interactions, quantifying the emergent effects of introduction of cold-chain facilities into the model.

Using the model can help answer a range of questions of such a system, including:

  • The location and prioritisation of investment into building cold-chain infrastructure within country to achieve mission-led targets such as economic growth or better population nourishment outcomes.
  • What the investment, staffing and vehicle requirements would be of such a system to deliver a reliable cold supply chain from smallholder farmers to consumers.
  • The effects of opening up new import sources or export points, and how geographical differences between regions manifest in the model’s response to such interventions.
  • Consideration of different modes of transport for delivery to market and how this affects the transport topology and the decisions based on it, such as the effects of opening a new railway line, or using delivery drones to bridge difficult-to-reach areas.
  • Analysing the effects of the “off-season” versus the harvest season, and showing how introduction of cooling hubs can help smooth out the peaked supply of perishable goods, resulting in greater predictability in prices.
  • Testing the model for resilience against disruption caused by extreme weather events, such as flooding blocking main arterial roads.
  • The effects of limiting range of deliveries over the road network to simulate adoption of battery electric delivery vehicles.
  • Testing how a climate-induced change in environmental conditions affects the strategy landscape and behaviour of the agents, such as higher temperatures affecting perishability of produce, or the effects of changing harvest conditions.

The key advantage of a virtual modelling environment is that it is designed to not just answer one specific question, but a class of questions, given accurate and representative input data, for example for the perishability and value effects of different produce, or the population demand landscape for such products, and how the demand changes over time. Virtual modelling also enables flexibility in the future design of novel systems, such as electric vehicles and charging stations.

Summary

The risks of global climate change affect the global population, and the importance of investing in the right facilities grows ever more pertinent to ensure resilience of food supply from Global South countries. Given the right data, agent-based virtual modelling is an important tool in cold-chain operations to allow practitioners and planners alike to analyse a wide range of potential investment and risk scenarios and gain a greater degree of assurance when making such decisions that deployment of cold-chain infrastructure will most effectively reduce waste, improve the country’s economy and improve the livelihoods of smallholder farmers.