⚡ Smart EV Charging & Grid Load Management System

A high-performance, concurrent EV Charging Station scheduling and load balancing engine built with Java 17 and Spring Boot.

The system solves two core challenges in EV infrastructure:

1. Zero Double-Booking Guarantee: Prevents race conditions during simultaneous slot bookings using an Augmented Interval Tree and fine-grained Lock Striping.
2. Grid-Aware Power Distribution: Dynamically allocates station power using a two-pass greedy load-balancing algorithm to prevent substation overloading.

---

🏛️ System Architecture

               +----------------------------------+
               |       EVStationController        |
               +-----------------+----------------+
                                 |
         +-----------------------+-----------------------+
         |                                               |
         v                                               v
+-------------------------------+             +-------------------------------+
| ConcurrentIntervalTreeService |             |    GridLoadBalancerService    |
| (Lock Striping per Bay)       |             | (Two-Pass Greedy Allocator)   |
+---------------+---------------+             +---------------+---------------+
                |                                             |
                v                                             v
+-------------------------------+             +-------------------------------+
|     IntervalTree (O(log N))   |             |   Active Charging Sessions    |
|  - Overlap Check: O(log N)    |             |   - Throttling (SoC >= 80%)   |
|  - Reservation Deletion       |             |   - Background Worker Pool    |
+-------------------------------+             +---------------+---------------+
                                                              |
                                                              v
                                              +-------------------------------+
                                              |       ChargingObserver        |
                                              | (Event-driven notifications)  |
                                              +-------------------------------+

---

🧠 Data Structures & Algorithms (DSA)

Why an Augmented Interval Tree?

In standard reservation systems, checking whether a time window $[t_{start}, t_{end}]$ overlaps with existing bookings requires an $O(N)$ linear scan across all reservations.

This project implements an Augmented Interval Tree (IntervalNode):

• Node Structure: Each node stores $[start, end]$, along with maxEnd (the maximum end time of any interval in its subtree).
• Subtree Pruning: During overlap searches, if the left subtree's maxEnd is earlier than the query's start time, the entire left branch is pruned in $O(1)$.
• Time Complexity:
  • Overlap Detection: $O(\log N)$ average case.
  • Insertion: $O(\log N)$ with bottom-up maxEnd maintenance.
  • Cancellation (Deletion): $O(\log N)$ node removal with subtree invariant restoration.

---

⚡ Concurrency & Multithreading

• Lock Striping: Instead of a single bottleneck lock across the entire station, each charging bay maintains its own ReentrantReadWriteLock. Parallel bookings for different bays execute concurrently without contention.
• Reader-Writer Separation: Non-blocking availability checks acquire readLock(), allowing unlimited simultaneous read queries. Slot reservations acquire writeLock().
• Stress-Tested Thread Safety: Verified through automated unit tests simulating 50 threads racing concurrently using CountDownLatch(1)—guaranteeing zero double-bookings.
• Background Worker Threads: Active charging simulations execute in an asynchronous ExecutorService pool without blocking HTTP request threads.

---

🎯 Design Patterns & OOP

• Observer Pattern (ChargingObserver): Decouples battery progression from user alerts. As charging sessions run in worker threads, events trigger when vehicles reach 80% (trickle charging) or 100% completion.
• Domain Modeling (ChargingBay & BayStatus): Clean separation of station infrastructure, tracking power capacity (e.g., 50 kW Fast DC vs. 22 kW AC) and operational states (AVAILABLE, RESERVED, CHARGING, MAINTENANCE).

---

🔌 REST API Endpoints

Method	Endpoint	Description
POST	/api/v1/ev/reserve	Reserve a time slot on a specific bay ($O(\log N)$ overlap check)
POST	/api/v1/ev/cancel	Cancel an existing reservation and free the slot
POST	/api/v1/ev/start-session	Launch an asynchronous charging session on a bay
GET	/api/v1/ev/bays	Inspect live status of all charging bays
GET	/api/v1/ev/grid-status	View real-time dynamic kW power allocation across vehicles

---

🛠️ Build & Run

Run Unit & Concurrency Tests:

./mvnw test

Start the Application:

./mvnw spring-boot:run

The application will start on http://localhost:8080.