Скачать или смотреть How to Override Values in ODE Integration Using SciPy's odeint

Learn effective strategies to override values during ODE integration with SciPy's `odeint`. Follow our guide to implement conditions and overcome integration challenges!
---
This video is based on the question https://stackoverflow.com/q/63496753/ asked by the user 'Hersh Bhargava' ( https://stackoverflow.com/u/399641/ ) and on the answer https://stackoverflow.com/a/63500871/ provided by the user 'Lutz Lehmann' ( https://stackoverflow.com/u/3088138/ ) at 'Stack Overflow' website. Thanks to these great users and Stackexchange community for their contributions.

Visit these links for original content and any more details, such as alternate solutions, latest updates/developments on topic, comments, revision history etc. For example, the original title of the Question was: Override values in ODE integration with scipy.integrate.odeint

Also, Content (except music) licensed under CC BY-SA https://meta.stackexchange.com/help/l...
The original Question post is licensed under the 'CC BY-SA 4.0' ( https://creativecommons.org/licenses/... ) license, and the original Answer post is licensed under the 'CC BY-SA 4.0' ( https://creativecommons.org/licenses/... ) license.

If anything seems off to you, please feel free to write me at vlogize [AT] gmail [DOT] com.
---
Overriding Values in ODE Integration with SciPy's odeint

When dealing with the numerical integration of systems of Ordinary Differential Equations (ODEs), you may encounter situations where you need to apply specific conditions to the output of your integration. A common challenge in this context is how to override computed values based on arbitrary criteria during the integration process. In this post, we will explore this issue and provide practical solutions using SciPy's odeint.

The Problem: Conditional Output Values

Imagine you are simulating a system of ODEs, and you have a requirement where if the value of a term (e.g., X[0]) falls below a certain threshold (let’s say 1), you want it to be set to 0 instead of using the computed value. Traditional numerical integration methods, such as scipy.integrate.odeint, treat the integration as a black box, where you do not have direct access to internal states or the ability to modify step values dynamically.

Example Scenario

Consider the following example where you want to integrate a toy ODE system defined by a function dXdt:

[[See Video to Reveal this Text or Code Snippet]]

The dXdt function returns derivatives based on the current state of the system. However, if X[0] drops below 1 during the computation, how can we replace this value with 0 during integration?

Current Workaround and its Limitations

One possible approach to achieve the desired behavior is to modify the dXdt function to react conditionally, as shown below:

[[See Video to Reveal this Text or Code Snippet]]

While this workaround modifies the integration path, it has clear drawbacks:

Non-generalizability: This approach is not adaptable to other conditions without further modification.

Approximation Issue: It does not provide an exact solution as it relies on manipulating the value in an oversimplified manner.

A Better Solution Using SciPy's solve_ivp

The odeint method lacks a mechanism to natively handle events or conditions on the go. Instead, consider using scipy.integrate.solve_ivp, which includes an event detection feature. Here's how you can leverage it:

Steps to Implement Event Detection

Define an Event Function: This function should identify when your condition occurs (e.g., crossing the threshold).

Use the solve_ivp Method: Integrate your ODE with the event detection included.

Modify the State Post-Event: Upon detecting the event, you can modify the state and restart the integration from that point.

For example:

[[See Video to Reveal this Text or Code Snippet]]

Conclusion

In conclusion, while odeint does not allow for direct manipulation of values during integration, switching to solve_ivp offers a practical approach by utilizing event detection mechanisms. This way, you can effectively override output values based on conditions, leading to more accurate and flexible integration results. Feel free to adapt this method to meet your specific needs in numerical simulations.

By leveraging these strategies, numerical integration can not only be iterative but also dynamic, responding to the evolving states of the system you're modeling.