Update README.md

README.md

@@ -15,9 +15,9 @@ tags:
 
 ## Overview
 
-- GreedRL is a Deep Reinforcement Learning (DRL) based solver that can solve various types of problems, such as TSP, VRPs (CVRP, VRPTW, VRPPD etc), Order Batching Problem, Knapsack Problem etc.
+- 🤠GreedRL is a Deep Reinforcement Learning (DRL) based solver that can solve various types of problems, such as TSP, VRPs (CVRP, VRPTW, VRPPD etc), Order Batching Problem, Knapsack Problem etc.
 
-- GreedRL achieves very high performance by running on GPU while generating high quality solutions.
+- 🤠GreedRL achieves very high performance by running on GPU while generating high quality solutions.
   **1200 times faster** than [Google OR-Tools](https://developers.google.com/optimization) for large-scale (>=1000 nodes) CVRP, and the solution quality is improved by **about 3%**.
 
 ## 🏆Award
@@ -27,7 +27,7 @@ tags:
 
 ## Editions
 
-We have deliveried the following two editions of GreedRL for users.
+We have deliveried the following two editions of 🤠GreedRL for users.
 
 - **The Community Edition** is open source and available to [download](https://huggingface.co/Cainiao-AI/GreedRL).
 - **The Enterprise Edition** has a higher performance implementation than **The Community Edition** (about 50 times faster), especially when solving larg-scale problems. For more informations, please contact <a href="mailto:[email protected]">us</a>.
@@ -38,8 +38,9 @@ We have deliveried the following two editions of GreedRL for users.
 
 ## COPs Modeling examples
 
+### Standard problems
 
-### Capacitated Vehicle Routing Problem (CVRP)
+#### Capacitated Vehicle Routing Problem (CVRP)
 <details>
     <summary>CVRP</summary>
 
@@ -94,7 +95,7 @@ class Objective:
 
 </details>
 
-### Pickup and Delivery Problem with Time Windows (PDPTW)
+#### Pickup and Delivery Problem with Time Windows (PDPTW)
 <details>
     <summary>PDPTW</summary>
 
@@ -175,7 +176,7 @@ class Objective:
 </details>
 
 
-### VRP with Time Windows (VRPTW)
+#### VRP with Time Windows (VRPTW)
 <details>
     <summary>VRPTW</summary>
 
@@ -257,7 +258,7 @@ class Objective:
 
 </details>
 
-### Travelling Salesman Problem (TSP)
+#### Travelling Salesman Problem (TSP)
 <details>
     <summary>TSP</summary>
 
@@ -306,7 +307,7 @@ class Objective:
 
 </details>
 
-### Split Delivery Vehicle Routing Problem (SDVRP)
+#### Split Delivery Vehicle Routing Problem (SDVRP)
 <details>
     <summary>SDVRP</summary>
 
@@ -358,9 +359,19 @@ class Objective:
 
 </details>
 
-### Instant Pickup and Delivery Service Problem
+### Real-world scenario problems
+
+In addition to being able to solve standard problems, 🤠GreedRL can also model and solve real-world scenario problems, like `Instant Pickup and Delivery Service` and `Order Batching Problem`.
+
+#### Instant Pickup and Delivery Service
+
+> Instant Pickup and Delivery Service are widespread in order dispatching systems of the supply chain, courier delivery services and elsewhere.
+> 
+> Orders are generated in real-time. A number of vehicles are scheduled to serve orders from pickup locations to delivery locations while respecting vehicle capacity and service time constraints. The objective is to dynamically assign each order to the most appropriate vehicle so that the overall transportation cost (e.g., overall distances) cound be minimized.
+
+
 <details>
-    <summary>Instant Pickup and Delivery Service Problem (PDP)</summary>
+    <summary>Instant Pickup and Delivery Service</summary>
 
 ```python
 from greedrl.feature import *
@@ -450,9 +461,9 @@ class Objective:
 
 </details>
 
-### Order Batching Problem
+#### Order Batching Problem
 <details>
-    <summary>Batching</summary>
+    <summary>Order Batching Problem</summary>
 
 ```python
 from greedrl import Problem, Solver