RV College of Go, change the World

Engineering ®

SLIME MOLD
ALGORITHM FOR
TRANSPORT NETWORKS
TEAM : IIT KENGERI

Advaith V R 1RV23BT009 Aditya K

1RV23AI007
Adithya A 1RV23AI006
Harshvardhan Y 1RV23ME052
Experiential learning
RV College of
Engineering ® Go, change the World

INTRODUCTION

● Slime mould algorithms are a fascinating class of bio-inspired optimization

techniques that mimic the foraging behavior of slime moulds, specifically their
ability to form efficient networks to locate and connect food sources.
● These algorithms excel in solving complex problems by leveraging
decentralized decision-making and adaptability.

● One of the most remarkable applications is in efficient city planning and

traffic optimization, where these algorithms help design urban layouts and
traffic systems that minimize congestion and improve overall flow, mirroring the
slime mould's innate ability to optimize its nutrient pathways.

Experiential learning
 RV College of
Engineering ® Go, change the World

PROJECT INSPIRATION

When a group of scientists in Japan

placed its favourite food, oat flakes, at
various points on the agar
corresponding to the major and
populous cities of Japan, this creature
was able to, within a timespan of 5-6
days, map out the complex rail network
which took Japanese engineers years to
plan.
This technology, perfected in nature,
can really help in traffic mapping in new
emerging countries in Africa and Asia
and city and housing development
planning.

Experiential learning
RV College of
Engineering ® Go, change the World

PROBLEM STATEMENT

• To develop a slime mold algorithm capable of optimizing

urban metro networks and to validate its effectiveness
through simulation and laboratory experiments.

• To test further test the effectiveness of the algorithm in

the application of navigation for using it for path-finding
to solve a maze

Experiential learning
RV College of
Engineering ® Go, change the World

Why Slime Mould Algorithms are the next big thing in the world of
computation?

• Conventional methods for solving the Travelling

Salesman Problem (TSP) include the brute force
method, which has a huge factorial time complexity is
infeasible for large numbers of cities, and dynamic
programming is impractical for very large instances.
• In contrast, the slime mould algorithm (SMA) offers
significant advantages: it adapts its network
dynamically, efficiently exploring and converging
towards optimal solutions while avoiding local optima,
and it uses decentralized decision-making, leading to
robust problem-solving and faster route discovery.
These characteristics make SMA more efficient and
scalable for complex optimization problems compared
to conventional algorithms. Experiential learning
RV College of
Engineering ® Go, change the World

Demonstrating Prototype of Algorithm in a recreation of the Japan

Metro Mould Experiment

Visit:

https://thecodeworks.in/sma/

Experiential learning
RV College of
Engineering ®
Go, change the World

SILK
BOARD

B’LORE
AIRPORT

Experiential learning
RV College of
Engineering ® Go, change the World

METHODOLOGY
1.Software Simulation Phase:
• Develop an agent-based model using the p5.js library.
• Agents mimic slime mold foraging behavior, moving toward predefined food sources (representing important
metro points).
• Fine-tune parameters (e.g., agent count, sensor offset, turn angle) for realistic behavior.
• Visualize agents’ paths to observe emergent network patterns.
2.Real-World Experiment:
• Place Physarum polycephalum in petri dishes with agar medium.
• Position oat flakes strategically (matching important points from the simulation).
• Incubate petri dishes under controlled conditions to promote slime mold growth and movement.
• Document network formation using time-lapse photography and microscopy.
3.Data Analysis and Comparison:
• Quantitatively analyze networks from both simulation and real-world experiments.
• Apply graph theory metrics (e.g., efficiency, redundancy, robustness).
• Compare results to validate the algorithm’s accuracy and effectiveness.
• Gain insights for potential improvements and practical applications.
Experiential learning
RV College of
Engineering ® Go, change the World

PRINCIPLES OF SLIME MOLD ALGORITHM

1.Agent-Based Modeling: SMA uses agent-based modeling, where each agent represents a
slime mold entity. These agents operate autonomously based on simple rules, leading to the
emergence of complex patterns.

2.Environment and Food Sensing: Agents sense their environment using sensors. These
sensors detect the presence of food or other agents by measuring the concentration of a trail
substance left by moving agents. The gradient in this trail substance guides agents toward areas
of interest.

3.Directional Updates: Agents adjust their direction based on sensory information. If an agent
detects a higher concentration of the trail substance in a particular direction, it adjusts its heading
accordingly. Random exploration prevents premature convergence.

4.Trail Laying and Following: As agents move, they leave behind a trail substance. This indirect
communication guides other agents toward promising regions. The trail decays over time,
ensuring adaptability to changing conditions.
Experiential learning
RV College of
Engineering ® Go, change the World

PRINCIPLES OF SLIME MOLD ALGORITHM

5.Food Detection and Prioritization: Agents prioritize movement toward

detected food sources. Upon reaching a food source, they increase the
concentration of the trail substance, reinforcing successful paths.

6.Phase-Based Learning: SMA alternates between two phases:

• Approach Food: Exploration phase where tendrils extend outward, guided
by fitness values and random exploration.
• Wrap Food: Exploitation phase where tendrils contract around promising
food sources, refining solutions.

7.Mathematical Model: SMA cyclically alternates between the approach

(search) food phase and the wrap (exploit) food phase. Initialization involves
treating the slime mold as distinct entities with calculated fitness.

Experiential learning
RV College of
Engineering ® Go, change the World

PRINCIPLES OF SLIME MOLD ALGORITHM

8.Odor Perception: Detection of food occurs through odor
perception, captured by an equation.

9.Emergent Behavior: Despite simple local interactions, SMA

exhibits emergent behavior, solving complex problems and forming
intricate patterns.

10.Application Domains: SMA has found applications in network

design, facility location, and scheduling.

Experiential learning
RV College of
Engineering ® Go, change the World

PHASE III
Conduction of Physical Verification Experiment: Incubation of Physarum
polycephalum in RVCE Biotech Lab

Experiential learning
RV College of
Engineering ® Go, change the World

Day-1 PICTURES OF EXPERIMENT CONDUCTED

Day-3
Day-2

Experiential learning
RV College of
Engineering ® Go, change the World

PICTURES OF EXPERIMENT CONDUCTED

Day-4
Experiential learning
 RV College of
Engineering ® Go, change the World

PICTURES OF EXPERIMENT CONDUCTED

Day-5

Day-6
Experiential learning
RV College of
Engineering ® Go, change the World

PICTURES OF EXPERIMENT CONDUCTED

Day-6
Experiential learning
 RV College of
Engineering ® Go, change the World

PICTURES OF EXPERIMENT CONDUCTED

Day-7

Experiential learning
RV College of
Engineering ® Go, change the World

RESULTS OF COMPARATIVE STUDY OF GROWTH OF

MOULD AND SOFTWARE SLIME MOULD ALGORITHM

Key Takeaway 1:
Trail Laying Behavior:
● Slime Mould: Exhibits more colonial trail laying, creating a
dense network of interconnected paths.
● Algorithm: Displays branch-specific trail laying, forming
distinct, separate branches

Experiential learning
RV College of
Engineering ® Go, change the World

RESULTS OF COMPARATIVE STUDY OF GROWTH OF

MOULD AND SOFTWARE SLIME MOULD ALGORITHM

Key Takeaway 2:
Branching Complexity:
● Slime Mould: Demonstrates a very high level of branching,
creating numerous offshoots from the main paths.
● Algorithm: Shows significantly less branching, resulting in a
simpler network structure.

Experiential learning
RV College of
Engineering ® Go, change the World

RESULTS OF COMPARATIVE STUDY OF GROWTH OF

MOULD AND SOFTWARE SLIME MOULD ALGORITHM

Key Takeaway 3:
Exploration Levels:
● Slime Mould: Engages in a high level of exploration,
continuously seeking new areas and food sources.
● Algorithm: Exhibits lower levels of exploration
compared to the slime mould.

Experiential learning
RV College of
Engineering ® Go, change the World

RESULTS OF COMPARATIVE STUDY OF GROWTH OF

MOULD AND SOFTWARE SLIME MOULD ALGORITHM
Key Takeaway 4:
Adaptation of Detection Parameters:
● Slime Mould: Continuously adapts its food detection
distance/range and turning angles based on the current
availability of food.
● Algorithm: Uses fixed parameters for food detection and turning
angles, which do not vary with time.

Experiential learning
RV College of
Engineering ® Go, change the World

RESULTS OF COMPARATIVE STUDY OF GROWTH OF

MOULD AND SOFTWARE SLIME MOULD ALGORITHM
Key Takeaway 5:
Valuation of Food Sources:

● Slime Mould: Does not treat all food sources equally, showing
preferences based on the food's current value or worth.
● Algorithm: Considers all food sources to have the same value,
leading to uniform treatment of resources.

Experiential learning
RV College of
Engineering ® Go, change the World

RESULTS OF COMPARATIVE STUDY OF GROWTH OF

MOULD AND SOFTWARE SLIME MOULD ALGORITHM

Key Takeaway 6:
Mortality and Movement:

● Slime Mould: Is not immortal and slowly dies off with

reducing food sources.
● Algorithm: Agents are immortal and continuously move,
regardless of food availability.

Experiential learning
RV College of
Engineering ® Go, change the World

RESULTS OF COMPARATIVE STUDY OF GROWTH OF

MOULD AND SOFTWARE SLIME MOULD ALGORITHM

Key Takeaway 7:
Stability vs. Shifting Behavior:
● Slime Mould: Stabilizes over time, forming a consistent
network once an optimal path is found.
● Algorithm: Is ever-shifting, constantly updating paths, which
can be both an advantage and a disadvantage.

Experiential learning
RV College of
Engineering ® Go, change the World

RESULTS OF COMPARATIVE STUDY OF GROWTH OF

MOULD AND SOFTWARE SLIME MOULD ALGORITHM
Key Takeaway 8:
Terrain Adaptation:
● Slime Mould: Adapts to varying heights and works on various
terrains to find the best route.
● Algorithm: Currently operates on a 2D surface only, lacking
terrain adaptation.

Experiential learning
RV College of
Engineering ® Go, change the World

RESULTS OF COMPARATIVE STUDY OF GROWTH OF

MOULD AND SOFTWARE SLIME MOULD ALGORITHM
Key Takeaway 9:
Obstacle Interaction:
● Slime Mould: Can detect, avoid, and interact with obstacles
in its environment.
● Algorithm: The latest version has been updated to detect
and avoid obstacles, addressing this limitation of earlier
versions.

Experiential learning
RV College of
Engineering ® Go, change the World

PHASE IV
Development of Slime Mould Algorithm v4.1 : Obstacle sensing
and avoiding for pathfinding applications

Experiential learning
RV College of
Engineering ® Go, change the World

Experiential learning
RV College of
Engineering ® Go, change the World

APPLICATIONS OF SLIME MOLD ALGORITHM

1. Efficient City Design: Our algorithm can optimize urban layouts by creating
efficient pathways and connections, reducing travel distances and enhancing
accessibility within cities.

2. Efficient Multipoint Navigation for Supply Chain Logistics: By dynamically

adjusting routes between multiple delivery points, our algorithm can improve
logistics efficiency, minimizing travel time and costs for supply chains and
postal services.

Experiential learning
RV College of
Engineering ® Go, change the World

APPLICATIONS OF SLIME MOLD ALGORITHM

3. Traffic Optimization: Implemented in navigation apps and

automated car services, our algorithm can reduce congestion
by identifying and adapting to the most efficient routes in real-
time.

4. Swarm Robotics and Decentralized Control of Micro

Robots: The algorithm enables decentralized coordination of
micro-robots, allowing them to work together efficiently for
tasks like search and rescue or environmental monitoring.

Experiential learning
RV College of
Engineering ® Go, change the World

APPLICATIONS OF SLIME MOLD ALGORITHM

5. Control and Movement of Life-

Saving Cellular Robotics: Our
algorithm can guide cellular robots
through complex environments to
deliver treatments or perform repairs,
enhancing the precision and
effectiveness of medical interventions.

Experiential learning
RV College of
Engineering ® Go, change the World

THANK YOU.

Experiential learning