This model aims to simulate Competition and Displacement of Online Interpersonal Communication Platforms process from a bottom-up angle. Individual interpersonal communication platform adoption and abandonment serve as the micro-foundation of the simulation model. The evolution mode of platform user online communication network determines how present platform users adjust their communication relationships as well as how new users join that network. This evolution mode together with innovations proposed by individual interpersonal communication platforms would also have impacts on the platform competition and displacement process and result by influencing individual platform adoption and abandonment behaviors. Three scenes were designed to simulate some common competition situations occurred in the past and current time, that two homogeneous interpersonal communication platforms competed with each other when this kind of platforms first came into the public eye, that a late entrant platform with a major innovation competed with the leading incumbent platform during the following days, as well as that both the leading incumbent and the late entrant continued to propose many small innovations to compete in recent days, respectively.
Initial parameters are as follows: n(Nmax in the paper), denotes the final node number of the online communication network node. mi (m in the paper), denotes the initial degree of those initial network nodes and new added nodes. pc(Pc in the paper), denotes the proportion of links to be removed and added in each epoch. pst(Pv in the paper), denotes the proportion of nodes with a viscosity to some platforms. comeintime(Ti in the paper), denotes the epoch when Platform 2 joins the market. pit(Pi in the paper), denotes the proportion of nodes adopting Platform 2 immediately at epoch comeintime(Ti). ct(Ct in the paper), denotes the Innovation Effective Period length. In Scene 2, There is only one major platform proposed by Platform 2, and ct describes that length. However, in Scene 3, Platform 2 and 1 will propose innovations alternately. And so, we set ct=10000 in simulation program, and every jtt epochs, we alter the innovation proposer from one platform to the other. Hence in this scene, jtt actually denotes the Innovation Effective Period length instead of ct.

Overview

The Weather model is a procedural generation model designed to create realistic daily weather data for socioecological simulations. It generates synthetic weather time series for solar radiation, temperature, and precipitation using algorithms based on sinusoidal and double logistic functions. The model incorporates stochastic variation to mimic unpredictable weather patterns and aims to provide realistic yet flexible weather inputs for exploring diverse climate scenarios.

The Weather model can be used independently or integrated into larger models, providing realistic weather patterns without extensive coding or data collection. It can be customized to meet specific requirements, enabling users to gain a better understanding of the underlying mechanisms and have greater confidence in their applications.

…

This model visualizes gradient descent optimization - the fundamental algorithm used to train neural networks and other machine learning models. Agents represent different optimization algorithms searching for the minimum of a loss landscape (the “error surface” that ML models try to minimize during training).

The model demonstrates how different optimizer types (SGD, Momentum with different parameters) behave on various loss landscapes, from simple bowls to the notoriously difficult Rosenbrock “banana valley” function. This helps build intuition about why certain optimization algorithms work better than others for different problem geometries.

HOW IT WORKS

…

This model is intended to study the way information is collectively managed (i.e. shared, collected, processed, and stored) in a system and how it performs during a crisis or disaster. Performance is assessed in terms of the system’s ability to provide the information needed to the actors who need it when they need it. There are two main types of actors in the simulation, namely communities and professional responders. Their ability to exchange information is crucial to improve the system’s performance as each of them has direct access to only part of the information they need.

In a nutshell, the following occurs during a simulation. Due to a disaster, a series of randomly occurring disruptive events takes place. The actors in the simulation need to keep track of such events. Specifically, each event generates information needs for the different actors, which increases the information gaps (i.e. the “piles” of unaddressed information needs). In order to reduce the information gaps, the actors need to “discover” the pieces of information they need. The desired behavior or performance of the system is to keep the information gaps as low as possible, which is to address as many information needs as possible as they occur.

The original Ache model is used to explore different distributions of resources on the landscape and it’s effect on optimal strategies of the camps on hunting and camp movement.

The MOBILITY model analyzes how agents’ mobility affects the performance of social-ecological systems in different landscape configurations.

Comparing impact of alternative behavioral theories in a simple social-ecological system.

A model of innovation diffusion in a structured population with two groups who are averse to adopting a produce popular with the outgroup.

The purpose of the AdaptPumpa model is to analyze the robustness of the Pumpa irrigation system in Nepal to climate change.

This model implements a classic scenario used in Reinforcement Learning problem, the “Cliff Walking Problem”. Consider the gridworld shown below (SUTTON; BARTO, 2018). This is a standard undiscounted, episodic task, with start and goal states, and the usual actions causing movement up, down, right, and left. Reward is -1 on all transitions except those into the region marked “The Cliff.” Stepping into this region incurs a reward of -100 and sends the agent instantly back to the start (SUTTON; BARTO, 2018).

The problem is solved in this model using the Q-Learning algorithm. The algorithm is implemented with the support of the NetLogo Q-Learning Extension